Wednesday, 2016-03-23

notmynamethe b7281cf commit is the earliest that works unchanged (because of the bind_port made required)00:05
*** hosanai has joined #openstack-swift00:05
notmynamehosanai: new nick?00:06
*** hosanai has quit IRC00:07
*** baojg has joined #openstack-swift00:07
kota_notmyname: hosanai quit :/00:07
*** hosanai has joined #openstack-swift00:07
kota_oh, join again.00:07
notmynameand is back now :-)00:07
hosanaigood morning! i changed nick from ho_. this nick name is registered one :-)00:08
notmynamegreat :-)00:08
*** baojg has quit IRC00:08
kota_hosanai: alright and you looks not in office today.00:09
*** ChanServ sets mode: +v hosanai00:09
hosanaikota_: yeah, i used different nick for private one but i think i can use this nick from home and office.00:11
kota_hosanai: got it00:12
kota_hosanai: and congrats for joining core team :)00:12
kota_we can see the members, here https://review.openstack.org/#/admin/groups/24,members00:13
kota_s/joining/joining to/00:13
hosanaikota_: hehe. thanks!00:14
*** daemontool has quit IRC00:15
kota_hosanai: I am happy, there are more core reviewers in Asian/Pacific Timezone :)00:15
hosanaikota_: yeah, i have been gradualy become nervous :-)00:17
kota_hosanai: nervous!? why?00:18
kota_hosanai: can i help you on something?00:19
hosanaikota_: kind of pressure. obligation to  behave as a core00:20
hosanaikota_: i will do step by step :-)00:20
*** rickyrem has joined #openstack-swift00:21
kota_hosanai: I was with, please ping me anytime if you need to help on something :)00:21
hosanaikota_: thanks!00:22
*** garthb has quit IRC00:22
*** mingdang1 has joined #openstack-swift00:25
*** mingdang1 has quit IRC00:31
claygH201!00:44
claygnotmyname: so we have always supported rolling-upgrades and our review team has a lot of experience thinking about and testing situations that arise in upgrade - so... surely we can work something out?  Did they give us you have until X to deliver Y?  Or where they just like oh btw lol!00:49
*** klamath has quit IRC00:50
*** NM has quit IRC00:58
notmynameclayg: they delayed approving the removal for one more week, since that was the first time I had seen it01:01
notmynameand torgomatic timburke and I talked about what to do today01:01
notmynamethe current thought it to use a really old version of the code and start either proxy or storage servers. then use the currently-being-tested patch to start the other one. then run functests01:03
notmynameso the timing is unfortunate. ideally, I'd love to see something in place by next week so that we don't have to worry about the whole foundation marketing that swift doesn't support rolling upgrades as part of the mitaka release01:04
notmynameIMO setting up the tests is the hard part. timburke's looking into that. how to set up two different versions of swift in one VM01:09
claygnotmyname: do *any* of the other projects do this?01:10
notmynamenova might. but which one of their tests actually does it wasn't obvious01:10
notmynameI assume nova does because they have the same tag and it wasn't marked for removal01:11
notmynameI'll be back online later tonight01:12
*** lyrrad has quit IRC01:16
*** StraubTW has joined #openstack-swift01:17
*** arch-nemesis has quit IRC01:24
*** MVenesio has joined #openstack-swift01:34
*** NM has joined #openstack-swift01:39
*** MVenesio has quit IRC01:39
*** baojg has joined #openstack-swift01:46
*** rickyrem has quit IRC01:54
*** panda has quit IRC02:05
*** panda has joined #openstack-swift02:05
jrichlihosanai: congrats on core!02:09
hosanaijrichli: thanks!02:10
*** htaccess has joined #openstack-swift02:15
htaccessis there a reccomended .net lib for swift?02:15
htaccessdon't see anything about swift in https://github.com/openstacknetsdk/openstack.net02:18
htaccessoh, wait: http://www.openstacknetsdk.org/docs/html/e11545c6-88c9-4ff1-b0cf-abffd4bd3ff7.htm02:19
*** baojg has quit IRC02:48
*** baojg has joined #openstack-swift02:50
*** haomaiwa_ has joined #openstack-swift02:51
*** haomaiwa_ has quit IRC03:01
*** haomaiwang has joined #openstack-swift03:01
*** mingdang1 has joined #openstack-swift03:04
*** nadeem has joined #openstack-swift03:09
*** dmorita has quit IRC03:12
notmynamegood evening03:16
* htaccess consults https://in-the-sky.org/twilightmap.php03:18
notmynamehtaccess: :-)03:19
*** StraubTW has quit IRC03:19
notmynamehtaccess: most of the active people are in north america, followed by some people in EU, then some in east asia/australia time zones. with several others scatter about from india to brazil to israel03:20
notmynamehtaccess: on http://docs.openstack.org/developer/swift/associated_projects.html#associated-projects I see SwiftBox (in C#) and openstack.net listed03:21
mahaticgood morning03:25
mahaticfrom india :)03:25
notmynamehello mahatic03:25
notmyname:-)03:25
mahaticsorry about brussels nonsense to everyone is staying in the region03:25
*** dmorita has joined #openstack-swift03:27
mahatichosanai: congrats on joining core team!03:27
htaccessnotmyname: thanks for the SwiftBox link (have a client looking to hook unity3d to swift)03:27
notmynamehtaccess: that's cool03:28
*** asettle has quit IRC03:28
*** links has joined #openstack-swift03:28
notmynamehtaccess: there are several game companies that have or are using swift. it would be cool to see more (yes, I'm making some assumptions there that unity3d==game)03:28
*** mingdang1 has quit IRC03:29
*** dmorita has quit IRC03:31
htaccessyea it broadly fits within the game area :)03:33
notmynamebroadly fits within the game area == people sortof have fun when they do it03:34
notmyname;-)03:34
htaccessit must be an important enough market ... amazon payed cytek for the privalage of forking their engine03:34
htaccessso they could give it away ... but require it uses aws of cloud setvices03:35
*** MVenesio has joined #openstack-swift03:35
*** NM has quit IRC03:35
notmynameah yes. the old "free* (* requires purchase of AWS)" plan03:36
*** MVenesio has quit IRC03:39
*** mingdang1 has joined #openstack-swift03:42
*** mingdang1 has quit IRC03:44
*** mingdang1 has joined #openstack-swift03:44
*** mingdang1 has quit IRC03:46
notmynameonly two more features to write up for the changelog: fast-post and concurrent gets03:47
*** arch-nemesis has joined #openstack-swift04:00
*** haomaiwang has quit IRC04:01
*** haomaiwang has joined #openstack-swift04:01
notmynamemattoliverau: I know you're not around, but thanks for your good commit message on the concurrent gets patch. it's very easy to fit into the changelog with very few changes :-)04:03
*** asettle has joined #openstack-swift04:06
*** arch-nemesis has quit IRC04:07
*** asettle has quit IRC04:10
*** haomaiw__ has joined #openstack-swift04:12
*** haomaiwang has quit IRC04:13
*** arch-nemesis has joined #openstack-swift04:20
*** treaki_ has joined #openstack-swift04:24
*** treaki__ has quit IRC04:26
notmynameyet again, writing up the changelog for a release reminds me of all the great stuff going into swift04:28
*** mrmoje has quit IRC04:35
openstackgerritKazuhiro MIYAHARA proposed openstack/swift: WIP: Swift Automated Tiering  https://review.openstack.org/28705704:38
claygi wonder where this came from -> Mar 23 04:51:27 saio swift: Failed to update sync_store /tmp/tmpFqe1bC/sda/containers/0/928/c0ad7ddcd21bd2c0303123f75421e928/c0ad7ddcd21bd2c0303123f75421e928.db: #012Traceback (most recent call last):#012  File "/vagrant/swift/swift/container/replicator.py", line 195, in _post_replicate_hook#012    self.sync_store.update_sync_store(broker)#012AttributeError: 'ContainerReplicator' object has n04:52
openstackgerritJohn Dickinson proposed openstack/swift: 2.7.0 authors and changelog updates  https://review.openstack.org/29617504:53
notmynamelots of good stuff in this releasee. please check the proposed changelog there for typos and accuracy04:53
notmynameif everything looks good, I'd like to land that tomorrow and start the release process so it can make it though the gate by friday04:53
*** haomaiw__ has quit IRC05:01
*** haomaiwang has joined #openstack-swift05:01
*** takashi has joined #openstack-swift05:18
*** StraubTW has joined #openstack-swift05:20
takashigood afternoon05:21
openstackgerritClay Gerrard proposed openstack/swift: try simpler setpriority/ioprio_set  https://review.openstack.org/29618105:21
*** StraubTW has quit IRC05:25
*** MVenesio has joined #openstack-swift05:36
*** MVenesio has quit IRC05:40
*** asettle has joined #openstack-swift05:55
*** haomaiwang has quit IRC06:01
*** haomaiwang has joined #openstack-swift06:01
*** panda has quit IRC06:05
*** panda has joined #openstack-swift06:06
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Imported Translations from Zanata  https://review.openstack.org/29619306:16
*** asettle has quit IRC06:25
*** ChubYann has quit IRC06:32
openstackgerritKazuhiro MIYAHARA proposed openstack/swift: WIP: Swift Automated Tiering  https://review.openstack.org/28705706:40
claygonovy: peterlisak: spent some time with nice patch 238799 today - great work there!06:41
patchbotclayg: https://review.openstack.org/#/c/238799/ - swift - Change schedule priority of daemon/server in config06:41
*** wanghua has joined #openstack-swift06:51
*** mingdang1 has joined #openstack-swift07:00
*** haomaiwang has quit IRC07:01
*** haomaiwang has joined #openstack-swift07:01
onovyclayg: cool, thanks for review07:06
*** haomaiwang has quit IRC07:06
openstackgerritClay Gerrard proposed openstack/swift: Auditor will clean up stale rsync tempfiles  https://review.openstack.org/29317707:09
*** haomaiwang has joined #openstack-swift07:09
*** mmcardle has quit IRC07:13
*** klrmn has quit IRC07:14
openstackgerritClay Gerrard proposed openstack/swift: Auditor will clean up stale rsync tempfiles  https://review.openstack.org/29317707:14
claygok, jeez I think i'm done messing with it for a few mins ^07:14
*** asettle has joined #openstack-swift07:18
*** asettle has quit IRC07:23
*** Trixboxer has joined #openstack-swift07:28
*** furlongm_ has joined #openstack-swift07:35
*** furlongm_ is now known as furlongm07:36
*** MVenesio has joined #openstack-swift07:37
furlongmhi, is it possible to set a different "default" storage policy on different proxy servers in the same global cluster?07:37
*** tesseract has joined #openstack-swift07:41
*** tesseract is now known as Guest6084207:41
*** MVenesio has quit IRC07:41
*** arch-nemesis has quit IRC07:50
*** asettle has joined #openstack-swift07:51
claygfurlongm: yeah you just have to distirbute different swift.conf's07:57
claygfurlongm: you could have every proxy have a different default storage policy!07:58
*** haomaiwang has quit IRC08:01
*** haomaiwa_ has joined #openstack-swift08:01
*** rledisez has joined #openstack-swift08:05
*** mrmoje has joined #openstack-swift08:07
*** swfituser has joined #openstack-swift08:08
swfituserhow can i use swift command to use a particular storage policy as against X-Storage-Policy using curl08:08
claygswiftuser: the swift commands can do arbitrary headers just like curl -H 'x-storage-policy: <name>'08:22
*** swfituser has quit IRC08:26
furlongmclayg: does the proxy server route based on the default storage policy in its swift.conf or on default storage policy on the storage nodes swift.conf. or both?08:27
*** asettle has quit IRC08:37
*** acoles_ is now known as acoles08:39
*** kei_yama has quit IRC08:41
*** asettle has joined #openstack-swift08:41
*** permalac has joined #openstack-swift08:43
*** jmccarthy has quit IRC08:49
*** jmccarthy has joined #openstack-swift08:50
acolesgood morning08:53
acolesclayg: you still awake? what was the context you saw that AttributeError in _post_replicate_hook? I can see how it *might* occur if _replicate_object is called directly08:54
acoleshosanai: timburke congrats!08:55
*** haomaiwa_ has quit IRC09:01
*** links has quit IRC09:01
*** haomaiwang has joined #openstack-swift09:01
*** stantonnet has quit IRC09:03
*** openstackgerrit has quit IRC09:03
*** openstackgerrit has joined #openstack-swift09:04
*** stantonnet has joined #openstack-swift09:05
mahaticacoles: good morning09:07
*** links has joined #openstack-swift09:17
acolesmahatic: hi09:22
*** StraubTW has joined #openstack-swift09:23
*** StraubTW has quit IRC09:27
*** takashi has quit IRC09:30
*** daemontool has joined #openstack-swift09:35
*** daemontool_ has joined #openstack-swift09:36
*** daemontool_ has quit IRC09:36
*** MVenesio has joined #openstack-swift09:37
*** MVenesio has quit IRC09:42
*** nadeem has quit IRC09:45
openstackgerritMerged openstack/swift: Imported Translations from Zanata  https://review.openstack.org/29619309:49
hosanaiacoles: hehe thanks :-)09:50
*** haomaiwang has quit IRC10:01
*** haomaiwang has joined #openstack-swift10:01
*** wanghua has quit IRC10:04
*** panda has quit IRC10:05
*** links has quit IRC10:05
*** baojg has quit IRC10:06
*** panda has joined #openstack-swift10:06
*** mmcardle has joined #openstack-swift10:09
*** mmcardle has quit IRC10:12
*** mmcardle has joined #openstack-swift10:13
*** mmcardle has joined #openstack-swift10:14
*** mmcardle has quit IRC10:14
*** hosanai has quit IRC10:17
*** links has joined #openstack-swift10:21
openstackgerritKazuhiro MIYAHARA proposed openstack/swift: WIP: Swift Automated Tiering  https://review.openstack.org/28705710:25
*** ankur has quit IRC10:31
*** mvk has joined #openstack-swift10:39
*** baojg has joined #openstack-swift10:42
*** furlongm_ has joined #openstack-swift10:44
*** furlongm has quit IRC10:45
*** baojg has quit IRC10:47
*** Jeffrey4l has joined #openstack-swift10:49
*** furlongm has joined #openstack-swift10:55
*** furlongm_ has quit IRC10:55
*** cbartz has joined #openstack-swift10:57
*** haomaiwang has quit IRC11:01
*** haomaiwang has joined #openstack-swift11:01
*** dRy has joined #openstack-swift11:04
*** Jeffrey4l has quit IRC11:10
*** Zyric has quit IRC11:14
*** Jeffrey4l has joined #openstack-swift11:17
*** Jeffrey4l has quit IRC11:28
openstackgerritMarek Kaleta proposed openstack/python-swiftclient: Fix SwiftPostObject options usage in SwiftService  https://review.openstack.org/29638211:30
*** jordanP has joined #openstack-swift11:36
*** MVenesio has joined #openstack-swift11:38
*** a1|away is now known as JelleB11:41
*** MVenesio has quit IRC11:43
*** Jeffrey4l has joined #openstack-swift11:47
*** john_bar has joined #openstack-swift11:54
*** silor has joined #openstack-swift11:54
*** joeljwright has joined #openstack-swift11:55
*** ChanServ sets mode: +v joeljwright11:55
*** Jeffrey4l has quit IRC11:58
*** david-lyle_ has joined #openstack-swift11:58
*** david-lyle has quit IRC12:00
*** haomaiwang has quit IRC12:01
*** haomaiwang has joined #openstack-swift12:01
*** joeljwright has quit IRC12:02
*** MVenesio has joined #openstack-swift12:02
*** ChanServ sets mode: +v tdasilva12:04
*** km has quit IRC12:07
*** silor1 has joined #openstack-swift12:10
*** silor has quit IRC12:11
*** silor1 is now known as silor12:11
*** dmorita has joined #openstack-swift12:13
*** haomaiwang has quit IRC12:16
*** dmorita has quit IRC12:17
*** asettle has quit IRC12:19
*** asettle has joined #openstack-swift12:20
*** takashi has joined #openstack-swift12:23
*** asettle has quit IRC12:24
*** Jeffrey4l has joined #openstack-swift12:25
*** mrmoje has quit IRC12:31
*** NM has joined #openstack-swift12:32
*** CaioBren_ has joined #openstack-swift12:32
*** CaioBrentano has quit IRC12:35
*** NM has quit IRC12:37
*** trifon has joined #openstack-swift12:37
*** Jeffrey4l has quit IRC12:38
*** NM has joined #openstack-swift12:40
*** Jeffrey4l has joined #openstack-swift12:40
*** StraubTW has joined #openstack-swift12:43
*** Jeffrey4l has quit IRC12:47
*** baojg has joined #openstack-swift12:50
*** haomaiwang has joined #openstack-swift12:57
openstackgerritTakashi Kajinami proposed openstack/swift: Use logger.exception when generating logs about exceptions  https://review.openstack.org/29642212:57
*** haomaiwang has quit IRC13:01
*** links has quit IRC13:01
*** david-lyle_ is now known as david-lyle13:06
*** MVenesio has quit IRC13:21
*** vint_bra has joined #openstack-swift13:24
*** daemontool_ has joined #openstack-swift13:33
*** daemontool has quit IRC13:33
peterlisakmattoliverau, acoles, kota_, clayg, Hi, thanks for your time, reviews and great comments in path #238799. I added my suggestion in last comment. Could you confirm it? (or refuse but I would prefer confirmation :))13:39
*** joeljwright has joined #openstack-swift13:39
*** ChanServ sets mode: +v joeljwright13:39
*** mmcardle has joined #openstack-swift13:39
peterlisakpath => patch #2387913:40
patchbotpeterlisak: https://review.openstack.org/#/c/23879/ - neutron - Adds support for the DHCP DNS domain search option... (ABANDONED)13:40
*** haomaiwang has joined #openstack-swift13:40
*** _JZ_ has joined #openstack-swift13:44
*** mvk_ has joined #openstack-swift13:45
*** mkrcmari__ has joined #openstack-swift13:48
*** mvk has quit IRC13:50
*** mvk_ has quit IRC13:52
*** mingdang1 has quit IRC13:53
*** mvk_ has joined #openstack-swift13:55
openstackgerritEran Rom proposed openstack/swift: Add process level concurrency to container sync  https://review.openstack.org/21009913:55
*** klamath has joined #openstack-swift13:56
klamathis there a way to create a storage policy for account/container?13:56
*** mvk has joined #openstack-swift13:57
*** mkrcmari__ has quit IRC13:58
*** BigWillie has joined #openstack-swift13:59
takashiklamath: you mean a storage policy for account data and container data, right?13:59
*** mvk_ has quit IRC13:59
takashiklamath: If it is right, no. Storage policiy is a feature only for object.14:00
*** haomaiwang has quit IRC14:01
*** haomaiwang has joined #openstack-swift14:01
*** mvk has quit IRC14:01
openstackgerritMarek Kaleta proposed openstack/python-swiftclient: Add copy object method  https://review.openstack.org/28020014:01
klamathok, thank you.  takashi14:01
*** panda has quit IRC14:04
takashiklamath: welcome :-)14:04
*** panda has joined #openstack-swift14:05
*** openstack has joined #openstack-swift14:23
*** ametts has joined #openstack-swift14:24
openstackgerritGleb Samsonov proposed openstack/swift: go: object-expirer  https://review.openstack.org/29648614:31
*** tongli has joined #openstack-swift14:34
openstackgerritMarek Kaleta proposed openstack/python-swiftclient: Fix SwiftPostObject options usage in SwiftService  https://review.openstack.org/29638214:38
*** david-lyle_ is now known as david-lyle14:47
*** Jeffrey4l has joined #openstack-swift14:52
*** arch-nemesis has joined #openstack-swift14:53
*** nadeem has joined #openstack-swift14:53
*** nadeem has quit IRC14:55
*** nadeem has joined #openstack-swift14:56
*** klrmn has joined #openstack-swift14:57
*** haomaiwang has quit IRC15:01
*** nadeem has quit IRC15:01
*** haomaiwang has joined #openstack-swift15:01
*** nadeem has joined #openstack-swift15:01
*** baojg has quit IRC15:03
notmynamegood morning15:04
openstackgerritMarek Kaleta proposed openstack/python-swiftclient: Add copy object method  https://review.openstack.org/28020015:05
notmynameacoles: thanks for the changelog comments15:07
*** cbartz has quit IRC15:09
*** garthb has joined #openstack-swift15:12
acolesnotmyname: np15:12
*** openstackstatus has joined #openstack-swift15:13
*** ChanServ sets mode: +v openstackstatus15:13
notmynameacoles: last night when putting that together, I noticed that the comments in the sample config weren't updated for post-as-copy. can you look if you get a couple of minutes? if you don't get to it, I'll proposed a patch later today https://github.com/openstack/swift/blob/master/etc/proxy-server.conf-sample#L13215:14
*** trifon has quit IRC15:16
acolesnotmyname: patch 29573915:20
patchbotacoles: https://review.openstack.org/#/c/295739/ - swift - Docs: Container sync does not require POST-as-COPY15:20
notmynamelol, ok. you're way ahead fo me :-)15:21
notmynameI'll look over that in a few minutes15:21
acolesnotmyname: just a few hours ;)15:21
notmynameoh, actually now. all my email is marked as read ;-)15:22
acolesi have a button that does that too15:22
*** mmcardle has joined #openstack-swift15:24
acolesclayg: if only you didn't write so many tests i could have your patch reviewed by now ;)15:25
*** gyee has joined #openstack-swift15:29
jrichliVersioning middleware docs say "The new object name (for the previous version) is <versions_container>/<length><object_name>/<timestamp>, where length is the 3-character zero-padded hexadecimal length of the <object_name> and <timestamp> is the timestamp of when the previous version was created."15:29
jrichliWhat if the max_object_name_length = 1024?  Is it documented somewhere that you may not be able to use the configured maximum object name length for versioned objects?  It seems to be capped at 999.  Or am I missing something?15:29
notmynamejrichli: nope, you're right. IIRC you'll end up with a shorter effective max length. I don't think it supersedes the constraint checks15:30
jrichliso, I guess the desire would be to specify more clearly in docs over changing behavior?  there could be an env flag sorta like we did in encryption if we wanted to change that.15:32
cschwedenotmyname: hello! is the wishlist for mitaka bugfixes still open? i now it’s late, but…15:33
cschwedepatch 257502 would be welcome in Mitaka ;)15:34
patchbotcschwede: https://review.openstack.org/#/c/257502/ - swift - Fix full_listing in internal_client15:34
*** tongli has quit IRC15:36
notmynameyeah, there's always going to be "one more thing" that would be nice in a release (for example, also patch 293177 and patch 270961)15:37
patchbotnotmyname: https://review.openstack.org/#/c/293177/ - swift - Auditor will clean up stale rsync tempfiles15:37
patchbotnotmyname: https://review.openstack.org/#/c/270961/ - swift - Container-Sync to perform HEAD before PUT object o...15:37
notmynameas of right now, we're in a good place for a release. as in, if we released with what's on master now (plus the authors/changelog update), I think we should all be pretty happy with it15:38
notmynameso, basically, I think anything else at this point is a nice to have, and not a release blocker. so yeah, if we can get reviews on it and have it land today (ie next ~12 hours), then great!15:39
cschwedepatch 293177 will be reviewed before todays meeting, i saw the updates from clayg on my comments15:39
patchbotcschwede: https://review.openstack.org/#/c/293177/ - swift - Auditor will clean up stale rsync tempfiles15:39
*** asettle has joined #openstack-swift15:39
notmynameok, https://wiki.openstack.org/wiki/Swift/PriorityReviews updated15:41
cschwedenotmyname: thx!15:41
notmynamejoeljwright just left a review! I thought he was on vacation15:42
notmynamejrichli: oh, good thought about EC fallocate15:42
notmynamejrichli: my first reaction is "sure", but then when I think of how to phrase it, it's one of those internal changes that's summarized as "things now properly work like you thought they always did"15:43
notmynameso maybe not?15:43
jrichlithat's why i asked too :-)15:43
jrichlii agree15:44
joeljwrightnotmyname: joeljwright is back15:44
notmynamewoohoo15:44
*** asettle has quit IRC15:44
joeljwrightalthough I seem to have scared asettle away...15:44
joeljwrightmaybe she knows I want to talk about docs again15:44
notmynameacoles: do you think the accept-ranges on EC responses needs to be called out specifically?15:45
acolesnotmyname: no string feeling15:46
notmynameacoles: any yarn feeling?15:46
acolesnotmyname: sorry, been a ropey day15:47
acolesdebugging thread15:47
notmynamelol15:47
notmynamejrichli: "Erasure coded fragments now properly call fallocate to reserve disk space before being written"15:47
notmyname?15:47
acolesnotmyname: fyi i am close to done on review of patch 29317715:48
patchbotacoles: https://review.openstack.org/#/c/293177/ - swift - Auditor will clean up stale rsync tempfiles15:48
notmynamenice. so between you and cschwede looks like we could get that one too15:48
claygi'm so lucky15:48
acolesclayg: heh15:48
jrichlinotmyname: sounds good to me!15:48
claygdid we ever land torgomatic's thing that cleaned up that terrible get_info mess?15:48
acolesi need to step back and decide how to vote15:48
claygacoles: if you're close enough that I bribe could help just let me know what I can do15:49
openstackgerritJohn Dickinson proposed openstack/swift: 2.7.0 authors and changelog updates  https://review.openstack.org/29617515:49
*** dRy has quit IRC15:49
claygj/k15:49
claygfurlongm: the proxy sets the default - it even sends it to the node as like x-backend-default-storage-policy-index or something crazy15:51
notmynameok, morning-tasks-while-drinking-coffee duties (email, reviews, IRC) are done. time to get ready to go to the office15:56
*** tongli has joined #openstack-swift15:56
*** mvk has joined #openstack-swift15:59
*** haomaiwang has quit IRC16:01
*** tongli has quit IRC16:01
*** haomaiwang has joined #openstack-swift16:01
*** links has joined #openstack-swift16:06
*** dmorita has joined #openstack-swift16:13
*** tongli has joined #openstack-swift16:17
claygcschwede: can you think of any other fallback that16:20
clayg... is more reasonable that reclaim age?16:20
*** Guest60842 has quit IRC16:21
claygcschwede: acoles had suggested it at some point - and it *is* longer that I would normally want - but definately long enough - i really don't think it should come up that often outside of tests where we manually hand in some {} that didn't come from readconf16:21
*** trifon has joined #openstack-swift16:22
*** tongli has quit IRC16:22
openstackgerritoshritf proposed openstack/swift: Container-Sync to perform HEAD before PUT object on remote  https://review.openstack.org/27096116:23
cschwedeclayg: well, i’m probably was more irritated by using reclaim_age and how it relates to this, but i understand that we just want to get a large fallback default value16:28
cschwedeclayg: but most likely no one except developers are reading this code part, so i’m ok with merging it like it is, if acoles is fine too16:28
*** klrmn has quit IRC16:29
claygacoles: "Time elapsed in seconds before rsync tempfiles will be unlinked. Config value of "auto" will use (in order of preference) object-replicator's rsync_timeout + 900, or reclaim_age, or 604800 (one week)." - when you spell it all out like that it just sounds stupid - sigh16:29
acolesclayg: cschwede : an alternative is to not perform the cleanup if there is no rsync_tempfile_timeout nor rsync_timeout available16:30
cschwedeok, looks like acoles put his +2 on it, so let’s merge16:30
claygcschwede: no acoles wants to spell it all out and make us look like jerks who can't help but write overly complicated code because we aren't smart enough to think of something better :'(16:30
cschwedeacoles: yeah, i was thinking to just use a static fallback default, but at the end the outcome is the same16:30
acolescschwede: like one week ;)16:31
*** nadeem has quit IRC16:31
cschwedewell, yes ;)16:31
clayg"Config value of "auto" will use object-replicator rsync_timeout + 900, of if not available cleanup will be disabled"16:31
claygthen you have to explain what "not available" means :\16:31
cschwedeso we keep it like it is?16:31
acolescschwede: i reasoned that reclaim_age should always be > rsync_timeout and is therefore 'safe' as a fallback16:32
claygcschwede: acoles had a lot of nits that should be cleaned up and a diff - i'm definately keen on spinning another one16:32
acolesclayg: we have 12 hours16:32
*** tongli has joined #openstack-swift16:32
claygbut... i know we have releases coming up and all - i think the auto config option is going to stay - i think the code works16:32
cschwedeyes, the auto config itself sounds like a good idea to me16:32
acoles+116:32
claygi mean I really like rsync_tempfile_timeout = [seconds|auto] - like I don't think we're going to break the interface16:32
claygit's just "wtf is auto" and how do we document it16:32
acolesclayg: s/auto/i'm_feeling_lucky/ :P16:33
claygrofl16:33
*** david-lyle has quit IRC16:33
*** tongli has quit IRC16:33
*** tongli has joined #openstack-swift16:33
*** tongli has quit IRC16:33
*** david-lyle has joined #openstack-swift16:33
claygyeah by the time you spell out what auto is acctually doing inthis patch you'd have to feel lucky to not just do the math yourself!16:33
acolesthen it self-documents16:33
clayg... which is also fine - ops are not stupid16:34
claygsorry - that's obvious of course - the joke was ment to be more like "ops aren't aren't cavalier and foolish like I am"16:35
claygacoles: cschwede: yeah so I think falling back to a second config option is too much16:35
ahaleidk which ops you are hanging out with !16:35
clayghow about "Config value of "auto" will use object-replicator rsync_timeout + 900, of if not available 86400"16:36
claygcschwede: acoles: ^16:36
openstackgerritEran Rom proposed openstack/swift: Add thread level concurrency to container sync  https://review.openstack.org/22533816:36
cschwedeclayg: sounds good to me16:37
acolescschwede: clayg ok16:37
claygNICE!16:38
claygso i'll fix the patch, you guys can patch my typos, and we can merge!16:39
* clayg will try to run spellcheck16:39
cschwedeclayg: yes, sounds like a plan for today16:39
*** lyrrad has joined #openstack-swift16:39
acolesclayg: for bonus points, add a note to rsync_tempfile_timeout docs to effect that "If not auto, this value should be greater than rsync_timeout" and vice-versa for rsync_timeout16:40
openstackgerritoshritf proposed openstack/swift: Per container stat. report  https://review.openstack.org/28181416:43
timburkegood morning16:45
*** links has quit IRC16:49
*** cdelatte has joined #openstack-swift16:53
*** tongli has joined #openstack-swift16:57
*** Zyric has joined #openstack-swift16:59
*** haomaiwang has quit IRC17:01
*** haomaiwang has joined #openstack-swift17:01
*** tongli has quit IRC17:01
*** htruta has joined #openstack-swift17:03
openstackgerritMerged openstack/swift: Docs: Container sync does not require POST-as-COPY  https://review.openstack.org/29573917:08
openstackgerritTim Burke proposed openstack/python-swiftclient: Follow-up to clean up an unnecessary variable  https://review.openstack.org/29662017:10
*** CaioBren_ has quit IRC17:17
*** CaioBrentano has joined #openstack-swift17:18
*** rledisez has quit IRC17:20
notmynamegood morning, again17:23
*** nadeem has joined #openstack-swift17:24
*** nadeem has quit IRC17:25
*** nadeem has joined #openstack-swift17:25
*** rcernin has quit IRC17:29
openstackgerritClay Gerrard proposed openstack/swift: Auditor will clean up stale rsync tempfiles  https://review.openstack.org/29317717:30
zigoHi guys. I have all of Mitaka rc1 validated in my tempest, however, we still got some Swift failures which I don't know how to fix. Could someone have a quick look into the log?17:30
zigohttps://mitaka-jessie.pkgs.mirantis.com/job/openstack-tempest-ci/54/consoleFull17:30
notmynamezigo: that is a *really* long page ;-)17:31
*** klrmn has joined #openstack-swift17:31
notmynameso I see some ceilometer error when scrolling up from the bottom17:31
zigonotmyname: search for tempest.api.object_storage17:32
zigoThere's many "The resource could not be found." errors.17:32
zigoProbably something isn't setup correctly.17:33
zigoFYI, this is the SAIO setup.17:33
zigo(with small changes/fixes)17:33
notmynamehmm17:33
notmynameso FWIW, all tests are green on this patch from this morning https://review.openstack.org/#/c/296175/17:33
patchbotnotmyname: patch 296175 - swift - 2.7.0 authors and changelog updates17:33
notmynamepoint is, if those are the same tempest tests, then there's likely a config error rather than a bug. maybe?17:33
claygacoles: this has never happened to me before - i'm not sure how I feel about it - you gave me a diff (https://gist.github.com/alistairncoles/8e7aeaf64780299a512e) and I acctually... I sorta acctually didn't... this is hard for me to say17:33
claygI liked the way I wrote it better :\17:34
zigonotmyname: That's what I expect, yeah !17:34
claygacoles: i... feel so ashamed17:34
zigoThough into *what* should I investigate?17:34
acolesclayg: no one can be perfect :P17:35
acolesclayg: note the ambiguity there ^^17:35
notmynamezigo: what the pipeline in your proxy config?17:35
zigonotmyname: pipeline = catch_errors gatekeeper healthcheck proxy-logging cache container_sync bulk tempurl ratelimit authtoken keystoneauth container-quotas account-quotas slo dlo versioned_writes proxy-logging proxy-server17:36
notmynamehmm...some of the tests failing are from staticweb, which you don't have in the pipeline. but some of those passed, too, so that's weird17:38
timburkeditto formpost17:38
notmynameyeah17:39
zigonotmyname: What is staticweb ?17:39
zigonotmyname: I can try to add it in the pipeline then...17:39
notmynamezigo: it's the one that does static site hosting in swift17:39
notmynamezigo: https://github.com/openstack/swift/blob/master/etc/proxy-server.conf-sample#L51917:39
claygacoles: rofl17:39
* zigo tries17:40
notmynamezigo: formpost is in the sample config at https://github.com/openstack/swift/blob/master/etc/proxy-server.conf-sample#L55717:40
timburkecontainer_sync may also need to have some extra config (looking at the "No realm key for 'realm1'" error); not sure exactly what that would look like, though17:40
claygnotmyname: it was probably negative tests - like this should 401 - yup totally 401's - nice work staticweb17:40
notmynamemaybe. the test name is ambiguous.  tempest.api.object_storage.test_container_staticweb.StaticWebTest.test_web_listing [0.063464s] ... ok17:41
zigonotmyname: Adding static web made  tempest.api.object_storage.test_container_staticweb.StaticWebTest.test_web_index work !!!17:43
notmynamezigo: timburke: I need to step away for a call17:43
notmynamezigo: great17:43
zigonotmyname: So likely, some other tests will also pass.17:43
zigoI'll fix this in my CI and do another run.17:43
zigoThanks so much.17:43
notmynamezigo: timburke: container sync is probably breaking because of the lack of https://github.com/openstack/swift/blob/master/etc/container-sync-realms.conf-sample17:43
zigonotmyname: What's formpost for?17:43
zigonotmyname: Should I also put it in the pipeline?17:44
timburkezigo: versioned_writes may need to be explicitly enabled; see https://github.com/openstack/swift/blob/master/etc/proxy-server.conf-sample#L75117:44
timburkezigo: formpost lets users upload objects directly to swift via a form on a webpage. as i recall, it uses a mechanism similar to tempurls (including the same keys)17:46
zigotimburke: Does this look correct?17:46
zigopipeline = catch_errors gatekeeper healthcheck proxy-logging cache container_sync bulk tempurl ratelimit formpost authtoken keystoneauth staticweb container-quotas account-quotas slo dlo versioned_writes proxy-logging proxy-server17:46
zigo(I have added formpost and staticweb, just not sure of the order...)17:47
timburkezigo: looks sane to me17:48
zigotimburke: tempest.api.object_storage.test_object_formpost.ObjectFormPostTest.test_post_object_using_form passes now! :)17:48
zigoGreat ! :)17:48
zigoLet's restart a CI run from scratch now.17:48
zigoThanks guys, that's really helpful.17:49
openstackgerritAlistair Coles proposed openstack/swift: Container-Sync to perform HEAD before PUT object on remote  https://review.openstack.org/27096117:51
*** mmcardle has quit IRC17:54
*** vinsh has quit IRC17:54
*** tongli has joined #openstack-swift17:54
*** vinsh has joined #openstack-swift17:54
*** tongli has quit IRC17:55
*** mrmoje has quit IRC17:55
*** tongli has joined #openstack-swift17:55
*** joeljwright has quit IRC17:57
*** jordanP has quit IRC18:00
*** haomaiwang has quit IRC18:01
*** haomaiwang has joined #openstack-swift18:01
*** vinsh has quit IRC18:01
*** vinsh has joined #openstack-swift18:02
acolesclayg: ^^ i just pushed a changed over oshrit's patch despite you totally destroying my self confidence :P I added +2 so if you do get chance to review it again then *maybe* we can squeeze it into the release18:03
acolesor if any other core is available to review ^^18:03
*** panda has quit IRC18:04
*** panda has joined #openstack-swift18:05
*** trifon has quit IRC18:06
claygso if the _object_in_remote_container method used x-newest - we'd avoid the 409/20218:16
*** mvk_ has joined #openstack-swift18:22
*** esker has joined #openstack-swift18:23
claygafk18:25
*** mvk has quit IRC18:25
*** arch-nemesis has quit IRC18:28
*** arch-nemesis has joined #openstack-swift18:35
*** permalac has quit IRC18:39
*** openstack has joined #openstack-swift19:22
*** zhiyan_ is now known as zhiyan19:23
*** serverascode_ is now known as serverascode19:23
*** openstackstatus has joined #openstack-swift19:24
*** ChanServ sets mode: +v openstackstatus19:24
*** mrmoje has joined #openstack-swift19:32
*** mrmoje has quit IRC19:33
*** awelleck has joined #openstack-swift19:36
claygjrichli: do you know Gleb Samsonov?19:36
claygI *think* he's an ibmer?19:36
jrichliclayg: doesn't ring a bell, but i can look him up19:36
openstackgerritAlistair Coles proposed openstack/swift: Auditor will clean up stale rsync tempfiles  https://review.openstack.org/29317719:40
acolestypos^^ all good19:40
claygacoles: lol19:44
acoleshey cschwede added +A :)19:45
*** ChubYann has joined #openstack-swift19:45
*** sheel has quit IRC19:47
acolesclayg: so my point about inconsistent configs was if a replicator config on node A has rsync_timeout=900, and it is pushing to node B but the replicator config on node B has rsync_timeout=10, then the auditor on B could gobble up "live" temp files. Crazy config, unlikely overlap of events, but could happen??19:49
claygacoles: oic19:55
claygacoles: yeah could totally happen hadn't considered a bunch of object-servers maybe running different rsync_timeout values - but i guess that was no real risk before19:55
claygacoles: i wonder if we talk about reclaim_age anywhere in the deployment guide to indicate it sorta has to be the same everywhere too19:56
claygacoles: another test might turn rsync_tempfile_timeout way down and see how bad the sending rsync client barfs about it?19:57
*** jmccarthy1 has quit IRC19:58
acolesclayg: seems like it could be good to have some place in docs that suggests using uniform values for those times19:58
acolesdisparate reclaim_ages could be interesting19:58
claygacoles: yeah I think the deployment guide would be nice19:59
*** haomaiw__ has quit IRC20:01
*** haomaiwang has joined #openstack-swift20:01
*** BigWillie has quit IRC20:13
*** openstack has joined #openstack-swift20:32
*** trifon has joined #openstack-swift20:37
*** chsc has joined #openstack-swift20:44
openstackgerritMerged openstack/swift: Container-Sync to perform HEAD before PUT object on remote  https://review.openstack.org/27096120:44
kota_good morning20:47
*** BigWillie has quit IRC20:48
*** chsc has quit IRC20:48
notmynamehello kota_20:49
kota_notmyname: hello20:49
*** chsc has joined #openstack-swift20:50
*** hosanai has joined #openstack-swift20:51
*** ChanServ sets mode: +v hosanai20:51
*** tongli has quit IRC20:58
*** takashi has joined #openstack-swift20:58
*** david-lyle has quit IRC20:59
notmynamemeeting time in #openstack-meeting21:00
*** haomaiwang has quit IRC21:01
*** haomaiwang has joined #openstack-swift21:01
notmynametorgomatic: clayg: ping21:01
openstackgerritAlistair Coles proposed openstack/swift: Check marker params in SimpleClient full listing requests  https://review.openstack.org/29680021:02
*** david-lyle has joined #openstack-swift21:05
*** silor has quit IRC21:07
*** m_kazuhiro has joined #openstack-swift21:09
*** m_kazuhiro has quit IRC21:15
*** m_kazuhiro has joined #openstack-swift21:18
*** david-lyle has quit IRC21:21
*** m_kazuhi_ has joined #openstack-swift21:22
*** david-lyle has joined #openstack-swift21:23
*** m_kazuhiro has quit IRC21:24
*** acoles is now known as acoles_21:25
*** acoles_ is now known as acoles21:25
*** acoles is now known as acoles_21:26
*** acoles_ is now known as acoles21:26
*** m_kazuhi_ has quit IRC21:26
*** m_kazuhiro has joined #openstack-swift21:30
*** m_kazuhi_ has joined #openstack-swift21:33
*** daemontool_ has quit IRC21:34
*** m_kazuhiro has quit IRC21:34
*** NM has quit IRC21:35
*** daemontool_ has joined #openstack-swift21:38
*** trifon has quit IRC21:38
*** asettle has joined #openstack-swift21:40
*** awelleck has left #openstack-swift21:40
claygoh did i miss it?21:41
timburkeyup :P21:41
claygnice21:41
notmyname"miss" it21:41
acolesnotmyname: my network was up and down during the discussion of rolling upgrade tests21:43
acolesnotmyname: but i think i got the gist21:43
notmynameacoles: you and cschwede will talk tomorrow morning your time while I'm asleep to see what's possible. then when I wake up, I'll ping you to see what we can do and we might talk on the phone21:44
acolesnotmyname: got it21:45
claygnotmyname: if you're doing a changelog update anyway - i acctually think the container-sync is worth a mention - i don't like the mythos that "container sync sucks" - it only sucks if we don't make it better - it's gotten a lot better as of late!21:45
notmynametrue :-)21:45
*** m_kazuhiro has joined #openstack-swift21:47
*** m_kazuhi_ has quit IRC21:48
acolesgood night21:49
*** m_kazuhiro has quit IRC21:53
openstackgerritChristian Schwede proposed openstack/swift: Ignore files in the devices directory when auditing objects  https://review.openstack.org/29518321:59
*** haomaiwang has quit IRC22:01
*** haomaiwang has joined #openstack-swift22:02
*** takashi has quit IRC22:02
*** hosanai has left #openstack-swift22:02
*** vint_bra has quit IRC22:04
*** panda has quit IRC22:04
*** panda has joined #openstack-swift22:05
*** trifon has joined #openstack-swift22:12
*** ametts has quit IRC22:14
notmynameclayg: https://gist.github.com/notmyname/abb7be32920b515a4c1d22:15
claygnotmyname: great!22:16
notmynameand https://gist.github.com/notmyname/a18c3803155b81c505cc22:19
openstackgerritMerged openstack/swift: Auditor will clean up stale rsync tempfiles  https://review.openstack.org/29317722:21
claygnotmyname: wfm - point is bigger number - better22:25
notmynameI've never seen that rule fail me (except for that time 5 days ago when it did)22:25
*** arch-nemesis has quit IRC22:26
openstackgerritJohn Dickinson proposed openstack/swift: 2.7.0 authors and changelog updates  https://review.openstack.org/29617522:26
openstackgerritJohn Dickinson proposed openstack/swift: 2.7.0 authors and changelog updates  https://review.openstack.org/29617522:26
*** sgundur has left #openstack-swift22:29
*** daemontool_ has quit IRC22:31
notmynamepatch 295543 looks interesting, as cross-openstack initiatives go22:32
patchbotnotmyname: https://review.openstack.org/#/c/295543/ - openstack-specs - Add centralized configuration options specification22:32
*** pgbridge has joined #openstack-swift22:33
*** acoles is now known as acoles_22:35
*** sheel has joined #openstack-swift22:40
*** nadeem has joined #openstack-swift22:40
*** tristanC has quit IRC22:43
*** garthb has joined #openstack-swift22:43
*** htaccess has left #openstack-swift22:43
*** tristanC has joined #openstack-swift22:44
*** nadeem has quit IRC22:49
*** Jeffrey4l has quit IRC22:49
*** NM has joined #openstack-swift22:51
*** NM has quit IRC22:53
*** chsc has quit IRC22:54
*** sgundur has joined #openstack-swift22:56
*** sgundur has left #openstack-swift22:56
*** acoles_ is now known as acoles22:59
acolescschwede: ping22:59
*** haomaiwang has quit IRC23:01
*** haomaiwa_ has joined #openstack-swift23:01
openstackgerritTim Burke proposed openstack/swift: Follow-up to patch 268723  https://review.openstack.org/29683123:02
*** km has joined #openstack-swift23:03
*** peterlisak has quit IRC23:04
*** darrenc_ has joined #openstack-swift23:05
*** darrenc has quit IRC23:06
*** mattoliverau has quit IRC23:06
*** darrenc_ is now known as darrenc23:07
*** matt6434 has joined #openstack-swift23:07
*** peterlisak has joined #openstack-swift23:11
*** BigWillie has joined #openstack-swift23:12
*** sgundur has joined #openstack-swift23:14
*** sgundur has left #openstack-swift23:14
openstackgerritAlistair Coles proposed openstack/swift: Check marker params in SimpleClient full listing requests  https://review.openstack.org/29680023:15
openstackgerritAlistair Coles proposed openstack/swift: Fix full_listing in internal_client  https://review.openstack.org/25750223:15
*** BigWillie has quit IRC23:16
*** trifon has quit IRC23:17
*** openstack has joined #openstack-swift23:24
acolesgood night again23:24
*** acoles is now known as acoles_23:25
*** daemontool_ has joined #openstack-swift23:26
*** daemontool_ has quit IRC23:27
*** kei_yama has joined #openstack-swift23:28

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