| *** catintheroof has joined #openstack-swift | 00:11 | |
| torgomatic | lespaul: you can run those curl commands that swift-get-nodes spits out and that'll tell you what's where, but AFAIK there's nothing to automate that for a single object | 00:15 |
|---|---|---|
| torgomatic | swift-dispersion-report does that for the entire cluster | 00:16 |
| *** jerrygb has quit IRC | 00:18 | |
| clayg | r with swift-dispersion-report or something similar you can always make the call to remove dead-not-coming-back nodes from the ring will reassign the primaries | 00:22 |
| clayg | ... | 00:23 |
| clayg | weird.. | 00:23 |
| clayg | i was trying to say that powering off the nodes won't trigger over-replication - the assumption is unavailability - not a loss in durability | 00:23 |
| clayg | if you start over-replication on split brain or network congestion or timeout you can fill up your cluster pretty quick | 00:24 |
| *** sams-gleb has joined #openstack-swift | 00:24 | |
| clayg | ... and then lastly yeah - if overall partition health dictates - you always have the option to make a ring change to remove the downed nodes and rebalance | 00:25 |
| *** admin6 has quit IRC | 00:29 | |
| *** sams-gleb has quit IRC | 00:29 | |
| *** jerrygb has joined #openstack-swift | 00:34 | |
| kota_ | morning | 00:40 |
| notmyname | if i'm reading the ML correctly, it looks like we need to tag a swiftclient release this week | 00:44 |
| *** PavelK has quit IRC | 00:46 | |
| clayg | notmyname: do you? i missed that? | 00:47 |
| notmyname | from http://lists.openstack.org/pipermail/openstack-dev/2017-January/110218.html | 00:48 |
| mattoliverau | notmyname: richard said this morning he needs to tag or organise a release for horizon this week, so yeah, we probably do. | 00:48 |
| notmyname | Library projects should be branched with, or shortly after, their | 00:49 |
| notmyname | final release this week (use Jan 19 as the deadline) | 00:49 |
| clayg | are we releasing Octocat this week? | 00:49 |
| notmyname | no | 00:50 |
| clayg | scared me | 00:50 |
| notmyname | but they always want the library projects to release quite a bit earlier | 00:50 |
| notmyname | oh! I think we have another week https://releases.openstack.org/ocata/schedule.html | 00:51 |
| notmyname | this is "non-client libraries" | 00:51 |
| mattoliverau | \o/ now get back to watching a talk or something :P | 00:52 |
| notmyname | anyone can watch https://linux.conf.au/stream | 00:54 |
| *** winggundamth has joined #openstack-swift | 01:05 | |
| *** catintheroof has quit IRC | 01:09 | |
| *** jerrygb has quit IRC | 01:13 | |
| *** winggundamth has quit IRC | 01:16 | |
| kota_ | clayg: here? | 01:17 |
| clayg | kota_: hi kota_ :D | 01:17 |
| kota_ | hi clayg | 01:17 |
| kota_ | I'm now looking/condidering about the patch https://review.openstack.org/#/c/418690/ | 01:17 |
| patchbot | patch 418690 - swift - Fix race in new partitions detecting new/invalid s... | 01:17 |
| kota_ | exactly, we should move to drop for using getmtime for file validation (i mean not modified since the last reference) | 01:19 |
| kota_ | clayg: but not sure I could follow your idea completely | 01:19 |
| kota_ | clayg: because the shell for-loop looks to report nothing you expected | 01:20 |
| clayg | kota_: my thought was just that that change needs to mock getmtime or else it doesn't pass reliably | 01:20 |
| kota_ | clayg: the console log is https://gist.github.com/bloodeagle40234/2239939c6cb42b6b23635936a427ef34 | 01:20 |
| clayg | kota_: oic, on my machine with your diff applied (to drop the getmtime mocking) it fails predicably for me | 01:20 |
| kota_ | clayg: it looks like, all tests passed for 10 times, correctly | 01:20 |
| clayg | the test works for me as written | 01:20 |
| clayg | ... only fails if I remove the getmtime mocking | 01:21 |
| kota_ | clayg: i ran it with my patch | 01:21 |
| clayg | perhaps your computer is slower than mine ;) or your mounted filesystem has better getmtime resolution? | 01:21 |
| kota_ | clayg: and then, I just checked also the situation Pavel reported (i.e. getmtime could increment the mtime in LXC), https://gist.github.com/bloodeagle40234/9eaf53aac466003cc8e6c43b0a0f66b2 | 01:21 |
| clayg | kota_: well I think that as long as you force it to be bigger the mock works | 01:22 |
| kota_ | it also doesn't fail, because even if we got different mtime there, IIRC, it's ok if no new entry comming to invalidation file? | 01:22 |
| clayg | it's only when the filesystem might return the same mtime for a file that was modified will you loose the second update | 01:22 |
| kota_ | clayg: ah, my laptop may be slower.... though :/ | 01:23 |
| *** trananhkma has quit IRC | 01:24 | |
| clayg | kota_: I believe it should be clear that an update can overwrite if the getmtime we take to start matches the getmtime we take before write_pickle - *but* somehow another process finished get_hashes while we were hashing and the *contents* of hashes.pkl are different (despite same mtime) | 01:24 |
| clayg | kota_: for me your patch to remove the getmtime mock failed like this -> https://gist.github.com/clayg/81be81ef339c7cc276b34c04be6c218c | 01:25 |
| kota_ | clayg: hooo, that looks worse, exactly | 01:26 |
| kota_ | ok it makes me sense. | 01:26 |
| kota_ | and if the patch 418690 land before patch 419787, patch 419787 will remove the mock (it doesn't look to remove yet) right? | 01:28 |
| patchbot | https://review.openstack.org/#/c/418690/ - swift - Fix race in new partitions detecting new/invalid s... | 01:28 |
| patchbot | https://review.openstack.org/#/c/419787/ - swift - Better optimistic lock in get_hashes | 01:28 |
| patchbot | https://review.openstack.org/#/c/419787/ - swift - Better optimistic lock in get_hashes | 01:28 |
| clayg | so OMM (and also apparently PavelK's) it is possible to have modified file return same getmtime value rarely - but rather than address that particular issue this change just mocks out getmtime so it can focus on only on not lossing the invalidate_suffix updates | 01:28 |
| kota_ | clayg: because the patch 419787 will replace the getmtime to inode info. | 01:28 |
| patchbot | https://review.openstack.org/#/c/419787/ - swift - Better optimistic lock in get_hashes | 01:28 |
| clayg | kota_: the patches are not currently "in series" | 01:29 |
| kota_ | clayg: ok, gotcha | 01:29 |
| kota_ | clayg: will turn out my score with final run-through | 01:30 |
| kota_ | clayg: thanks | 01:30 |
| clayg | if patch 418690 lands first - i'll probably have to rebase 419787 - it would probably be good to audit for no-longer-needed mocking of getmtime in test_diskfile - not sure if I did that | 01:30 |
| patchbot | https://review.openstack.org/#/c/418690/ - swift - Fix race in new partitions detecting new/invalid s... | 01:30 |
| clayg | maybe inside the mock - where it makes the call to create delete of the second suffix - if you check it's get_hashes result it probably finds/hashes both suffixes - so the problem is really only the earlier one that started stomping on the other processes' changes | 01:33 |
| clayg | yeah the inode thing is going to be a lot better for that | 01:33 |
| *** mingyu has joined #openstack-swift | 01:50 | |
| clayg | gah! I call set_replicas to double up for my ec duplication factor and then rebalance and get "The time between rebalances must be at least min_part_hours" but that's non-sense - you have pleanty of part-frags that you can reassign swift-ring-builder - we need patch 326967 | 01:53 |
| patchbot | https://review.openstack.org/#/c/326967/ - swift - Rebalance with min_part_seconds_left > 0 | 01:53 |
| clayg | timburke: first try!!!!!! swift info --json | jq .swift.policies[].name | 01:56 |
| clayg | oh look out! | 01:58 |
| clayg | https://gist.github.com/clayg/fefdf585aa1f570bf0fc57407f4cd4f1 | 01:58 |
| clayg | oh hrm... nothing really makes the duplicated frags avoid each others failure domains | 02:02 |
| *** jerrygb has joined #openstack-swift | 02:14 | |
| kota_ | clayg: oh sorry, I was apart a bit for being asked from my collegue, will take a look https://review.openstack.org/#/c/326967/ too. | 02:16 |
| patchbot | patch 326967 - swift - Rebalance with min_part_seconds_left > 0 | 02:16 |
| clayg | kota_: i was just griping in general | 02:17 |
| clayg | kota_: you don't have to reoganize your review backlog on my account - i'm over it | 02:17 |
| clayg | rebalance -f does the trick | 02:17 |
| kota_ | sure, that is a trick to force to rebalance | 02:18 |
| *** klrmn has quit IRC | 02:21 | |
| *** jamielennox is now known as jamielennox|away | 02:25 | |
| *** sams-gleb has joined #openstack-swift | 02:27 | |
| *** sams-gleb has quit IRC | 02:31 | |
| *** tqtran has quit IRC | 02:33 | |
| *** mingyu has quit IRC | 02:40 | |
| *** mingyu has joined #openstack-swift | 02:45 | |
| *** mingyu_ has joined #openstack-swift | 02:47 | |
| *** mingyu has quit IRC | 02:47 | |
| clayg | kota_: I left some notes on the ec duplicate frag patch - don't feel an particular need to address anything specificially if it strikes you as not quite correct - my review is a wip | 02:55 |
| clayg | dinnertime! | 02:55 |
| kota_ | clayg: much appreciated! enjoy your time! | 02:55 |
| *** JimCheung has quit IRC | 03:01 | |
| *** mingyu_ has quit IRC | 03:02 | |
| *** jamielennox|away is now known as jamielennox | 03:03 | |
| *** bkopilov has quit IRC | 03:05 | |
| *** bkopilov_ has quit IRC | 03:06 | |
| *** klrmn has joined #openstack-swift | 03:17 | |
| *** mingyu has joined #openstack-swift | 03:19 | |
| *** links has joined #openstack-swift | 03:32 | |
| *** klrmn has quit IRC | 03:42 | |
| *** mingyu has quit IRC | 03:48 | |
| *** mingyu has joined #openstack-swift | 03:54 | |
| *** takashi has joined #openstack-swift | 03:59 | |
| *** psachin has joined #openstack-swift | 04:09 | |
| *** sheel has joined #openstack-swift | 04:13 | |
| timburke | clayg: w00t! | 04:18 |
| *** jlwhite has quit IRC | 04:21 | |
| *** dmorita has quit IRC | 04:25 | |
| *** bkopilov has joined #openstack-swift | 04:27 | |
| *** mingyu has quit IRC | 04:27 | |
| *** mingyu has joined #openstack-swift | 04:27 | |
| *** mingyu has quit IRC | 04:27 | |
| *** sams-gleb has joined #openstack-swift | 04:29 | |
| *** sams-gleb has quit IRC | 04:34 | |
| *** bkopilov_ has joined #openstack-swift | 04:34 | |
| mahatic | clayg: thanks for taking care of this patch 390901! | 04:35 |
| patchbot | https://review.openstack.org/#/c/390901/ - swift - Confirm receipt of SLO PUT with etag (MERGED) | 04:35 |
| *** dmorita has joined #openstack-swift | 04:49 | |
| *** dmorita has quit IRC | 04:53 | |
| *** SkyRocknRoll has joined #openstack-swift | 04:54 | |
| *** ppai has joined #openstack-swift | 05:07 | |
| *** jerrygb has quit IRC | 05:08 | |
| *** jlwhite has joined #openstack-swift | 05:20 | |
| *** dfflanders has quit IRC | 05:26 | |
| *** Jeffrey4l_ has joined #openstack-swift | 06:30 | |
| *** sams-gleb has joined #openstack-swift | 06:31 | |
| *** tqtran has joined #openstack-swift | 06:33 | |
| *** Jeffrey4l has quit IRC | 06:34 | |
| *** Jeffrey4l_ is now known as Jeffrey4l | 06:34 | |
| *** tqtran has quit IRC | 06:35 | |
| *** sams-gleb has quit IRC | 06:36 | |
| *** ppai has quit IRC | 06:45 | |
| *** ppai has joined #openstack-swift | 07:03 | |
| *** tesseract has joined #openstack-swift | 07:16 | |
| *** ChubYann has quit IRC | 07:32 | |
| *** ppai has quit IRC | 07:49 | |
| *** dmellado has quit IRC | 08:02 | |
| *** ppai has joined #openstack-swift | 08:02 | |
| *** dmellado has joined #openstack-swift | 08:04 | |
| *** hseipp has joined #openstack-swift | 08:04 | |
| *** rledisez has joined #openstack-swift | 08:10 | |
| openstackgerrit | Mathias Bjoerkqvist proposed openstack/swift: Retrieve encryption root secret from Barbican https://review.openstack.org/364878 | 08:18 |
| *** geaaru has joined #openstack-swift | 08:23 | |
| *** sams-gleb has joined #openstack-swift | 08:33 | |
| *** sams-gleb has quit IRC | 08:38 | |
| *** ppai has quit IRC | 08:44 | |
| *** ppai has joined #openstack-swift | 09:01 | |
| *** sams-gleb has joined #openstack-swift | 09:14 | |
| *** mvk has quit IRC | 09:30 | |
| *** Shashikant86 has joined #openstack-swift | 09:32 | |
| *** cbartz has joined #openstack-swift | 09:36 | |
| *** Shashikant86 has quit IRC | 09:37 | |
| *** acoles_ is now known as acoles | 09:38 | |
| *** jordanP has joined #openstack-swift | 09:38 | |
| *** Shashikant86 has joined #openstack-swift | 09:45 | |
| acoles | clayg: lol @ " maybe folks are scared to review the other suffix fixes because they think I get frustrated things are moving to slowly and go +A you patch anyway (I totaly might!)" so yeah, which way do i bet? :) | 09:46 |
| acoles | PavelK (ah, he's not here) fwiw I am going to review clayg's set of patches | 09:48 |
| *** hseipp has quit IRC | 09:50 | |
| *** hseipp has joined #openstack-swift | 09:54 | |
| *** jordanP has quit IRC | 10:04 | |
| *** mvk has joined #openstack-swift | 10:06 | |
| *** Shashikant86 has quit IRC | 10:19 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: EC Fragment Duplication - Foundational Global EC Cluster Support https://review.openstack.org/219165 | 10:27 |
| *** kei_yama has quit IRC | 10:29 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Add support to increase object ring partition power https://review.openstack.org/337297 | 10:29 |
| *** mvk has quit IRC | 10:32 | |
| *** mvk has joined #openstack-swift | 10:47 | |
| *** sams-gleb has quit IRC | 10:50 | |
| *** sams-gleb has joined #openstack-swift | 10:50 | |
| *** mingyu has joined #openstack-swift | 10:53 | |
| mathiasb | what's the easiest way of figuring out where a 401 error originates from? | 10:54 |
| mathiasb | nothing obvious in /var/log/swift/; swift stat <cont> <obj> succeeds; swift stat <cont> fails | 10:55 |
| *** sams-gleb has quit IRC | 10:55 | |
| *** Shashikant86 has joined #openstack-swift | 11:02 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Add support to increase object ring partition power https://review.openstack.org/337297 | 11:08 |
| *** sams-gleb has joined #openstack-swift | 11:09 | |
| *** takashi has quit IRC | 11:10 | |
| *** mingyu has quit IRC | 11:15 | |
| *** oshritf_ has joined #openstack-swift | 11:17 | |
| *** hseipp has quit IRC | 11:19 | |
| *** hseipp has joined #openstack-swift | 11:20 | |
| *** jordanP has joined #openstack-swift | 11:25 | |
| *** dmorita has joined #openstack-swift | 11:26 | |
| *** mvk has quit IRC | 11:28 | |
| *** dmorita has quit IRC | 11:31 | |
| *** sams-gle_ has joined #openstack-swift | 11:36 | |
| *** sams-gleb has quit IRC | 11:36 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Rebalance with min_part_seconds_left > 0 https://review.openstack.org/326967 | 11:37 |
| *** ganders has joined #openstack-swift | 11:40 | |
| *** mvk has joined #openstack-swift | 11:41 | |
| *** ppai has quit IRC | 11:41 | |
| *** ppai has joined #openstack-swift | 11:42 | |
| *** cdelatte has joined #openstack-swift | 11:59 | |
| *** links has quit IRC | 12:12 | |
| *** Shashikant86 has quit IRC | 12:13 | |
| *** bkopilov_ has quit IRC | 12:22 | |
| *** bkopilov has quit IRC | 12:23 | |
| *** tesseract has quit IRC | 12:23 | |
| *** catintheroof has joined #openstack-swift | 12:32 | |
| *** ppai has quit IRC | 12:44 | |
| *** klamath has joined #openstack-swift | 12:44 | |
| *** klamath has quit IRC | 12:45 | |
| *** klamath has joined #openstack-swift | 12:45 | |
| acoles | mathiasb: not sure but could be if you have a .r:* ACL on the container. I *think* that 401 (vs 403) implies not token/identity supplied with the request | 12:49 |
| * acoles afk | 12:49 | |
| *** SkyRocknRoll has quit IRC | 12:53 | |
| *** ppai has joined #openstack-swift | 12:57 | |
| *** links has joined #openstack-swift | 12:58 | |
| *** Shashikant86 has joined #openstack-swift | 12:58 | |
| *** links has quit IRC | 13:01 | |
| *** vint_bra has joined #openstack-swift | 13:13 | |
| *** Shashikant86 has quit IRC | 13:29 | |
| mathiasb | acoles: thanks, I tried changing the ACL but no luck with the GET/stat so far | 13:45 |
| *** mingyu has joined #openstack-swift | 13:45 | |
| *** ppai has quit IRC | 13:50 | |
| *** Shashikant86 has joined #openstack-swift | 14:00 | |
| *** Shashikant86 has quit IRC | 14:00 | |
| acoles | mathiasb: if this is keystone auth then if you set the proxy log level to debug you may see some logging that indicates the basis on which the object GET is granted. Also look for any message in of the 401 response. | 14:09 |
| acoles | s/in of/in body of/ | 14:10 |
| *** SkyRocknRoll has joined #openstack-swift | 14:14 | |
| *** jerrygb has joined #openstack-swift | 14:14 | |
| *** Shashikant86 has joined #openstack-swift | 14:15 | |
| *** sams-gle_ has quit IRC | 14:19 | |
| *** bkopilov_ has joined #openstack-swift | 14:32 | |
| *** bkopilov has joined #openstack-swift | 14:33 | |
| *** mingyu has quit IRC | 14:34 | |
| *** sams-gleb has joined #openstack-swift | 14:36 | |
| *** tqtran has joined #openstack-swift | 14:36 | |
| *** psachin has quit IRC | 14:37 | |
| *** tqtran has quit IRC | 14:38 | |
| *** SkyRocknRoll has quit IRC | 14:43 | |
| *** mingyu has joined #openstack-swift | 14:49 | |
| *** manij has quit IRC | 15:00 | |
| *** manij has joined #openstack-swift | 15:01 | |
| *** manij has quit IRC | 15:05 | |
| *** mingyu has quit IRC | 15:08 | |
| *** mingyu has joined #openstack-swift | 15:09 | |
| *** pcaruana has quit IRC | 15:15 | |
| *** david-lyle has joined #openstack-swift | 15:42 | |
| *** mvk has quit IRC | 15:51 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Pretend *some* parts min_part_hours_passed https://review.openstack.org/311226 | 15:51 |
| *** tanee has quit IRC | 15:53 | |
| *** bikmak has joined #openstack-swift | 15:53 | |
| *** david-lyle has quit IRC | 15:55 | |
| *** _JZ_ has joined #openstack-swift | 15:58 | |
| *** mingyu has quit IRC | 16:03 | |
| *** Shashikant86 has quit IRC | 16:06 | |
| *** sheel has quit IRC | 16:07 | |
| *** david-lyle has joined #openstack-swift | 16:07 | |
| *** Shashikant86 has joined #openstack-swift | 16:10 | |
| *** chsc has joined #openstack-swift | 16:35 | |
| *** Shashikant86 has quit IRC | 16:43 | |
| *** Shashikant86 has joined #openstack-swift | 16:46 | |
| *** tanee has joined #openstack-swift | 16:46 | |
| *** oshritf_ has quit IRC | 16:54 | |
| *** silor has joined #openstack-swift | 17:00 | |
| *** Shashikant86 has quit IRC | 17:03 | |
| *** ganders has quit IRC | 17:04 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Add more tests for diskfile consolidate_hashes https://review.openstack.org/421385 | 17:08 |
| acoles | clayg: fyi I am on patch 418691 but got distracted by untested conditions | 17:09 |
| patchbot | https://review.openstack.org/#/c/418691/ - swift - Fix performance regression with hash invalidations | 17:09 |
| *** mvk has joined #openstack-swift | 17:17 | |
| *** geaaru has quit IRC | 17:24 | |
| *** hseipp has quit IRC | 17:28 | |
| *** ganders has joined #openstack-swift | 17:30 | |
| *** vinsh has joined #openstack-swift | 17:30 | |
| timburke | good morning | 17:43 |
| *** dmorita has joined #openstack-swift | 17:44 | |
| *** rledisez has quit IRC | 17:52 | |
| *** rledisez has joined #openstack-swift | 17:56 | |
| *** rledisez has quit IRC | 17:57 | |
| *** rledisez has joined #openstack-swift | 17:59 | |
| *** rledisez has quit IRC | 18:00 | |
| cbartz | timburke: Good morning. Is it possible that you could take a look over patch 414232 ? | 18:11 |
| patchbot | https://review.openstack.org/#/c/414232/ - swift - Exclude containers for account quota | 18:11 |
| *** ganders has quit IRC | 18:30 | |
| *** ganders has joined #openstack-swift | 18:34 | |
| *** tqtran has joined #openstack-swift | 18:38 | |
| *** klrmn has joined #openstack-swift | 18:39 | |
| *** tqtran has quit IRC | 18:40 | |
| *** david-lyle has quit IRC | 18:47 | |
| *** JimCheung has joined #openstack-swift | 18:48 | |
| *** tqtran has joined #openstack-swift | 19:04 | |
| *** acoles is now known as acoles_ | 19:07 | |
| *** klrmn has quit IRC | 19:24 | |
| *** klrmn has joined #openstack-swift | 19:28 | |
| *** klrmn has quit IRC | 19:28 | |
| *** PavelK has joined #openstack-swift | 19:31 | |
| *** klrmn has joined #openstack-swift | 19:34 | |
| *** ChubYann has joined #openstack-swift | 19:37 | |
| *** lespaul has quit IRC | 19:40 | |
| *** tqtran has quit IRC | 19:41 | |
| *** NM has joined #openstack-swift | 19:47 | |
| *** ganders has quit IRC | 19:48 | |
| *** jordanP has quit IRC | 19:49 | |
| *** Fin1te has joined #openstack-swift | 19:49 | |
| *** david-lyle has joined #openstack-swift | 19:50 | |
| *** Fin1te has left #openstack-swift | 19:50 | |
| caiobrentano_ | Hi folks... does anyone knows about SLO objects? | 19:51 |
| timburke | i love SLOs! | 19:52 |
| caiobrentano_ | Checking my logs, I'm seeing repeated PUT requests (from the SLO middleware), for the same segments | 19:52 |
| caiobrentano_ | timburke: lol | 19:52 |
| caiobrentano_ | is this the expected behaviour? | 19:53 |
| timburke | it's lying, it's doing HEADs against the segments (to verify they exist, have the correct size/etag if provided, ...) during the PUT for the manifest | 19:53 |
| caiobrentano_ | so, the method is wrong in the log? | 19:54 |
| timburke | at least, as i recall. we should really fix that. you should be able to confirm that's what's going on by looking at the object-server logs for the same transaction | 19:54 |
| timburke | yup | 19:54 |
| timburke | i mean, kind of. the client method *is* a PUT, but that's not what's going on with the backend | 19:54 |
| caiobrentano_ | my log record: http://pastebin.com/1UDisbKC | 19:56 |
| *** silor has quit IRC | 19:57 | |
| caiobrentano_ | timburke I don't think I understood what you said about "the client method *is* a PUT, but that's not what's going on with the backend" | 19:58 |
| *** tqtran has joined #openstack-swift | 20:03 | |
| timburke | when the client wants to complete the SLO, they send a PUT request with a JSON body listing all of the segments that should be used. during that PUT, the SLO middleware issues a bunch of HEAD requests to the segments before finally sending the PUT for the manifest. so all these things happen as a result of the external client sending a PUT, but that isn't terribly informative or useful so it'd be better if we logged it as a HEAD (whic | 20:04 |
| timburke | h it was) | 20:04 |
| *** tqtran has quit IRC | 20:05 | |
| timburke | caiobrentano_: from your logs, that first line (tx34dfb28714b1406086c0f) looks like the actual upload of the segment. the subsequent lines were likely the result of manifest PUTs and should have been recorded as HEADs. you can confirm that by grepping for (for example) tx8cd5c665d17b4e689ceaf in your object-server logs and checking the request method there | 20:07 |
| caiobrentano_ | timburke got it! thanks for the class about SLO! :) | 20:07 |
| caiobrentano_ | timburke: I did this grep on the object-server I found the HEAD request | 20:08 |
| NM | timburke: Caio has left the building but thanks again. By any reason could SLO delete a segment? | 20:18 |
| timburke | NM: if the client sends a DELETE to the manifest with ?multipart-manifest=delete, then the SLO middleware will delete all of the segments as well | 20:19 |
| NM | We couldn't find any DELETE request. But we found a .ts file where the segment should be. | 20:39 |
| timburke | so *someone* deleted them. may have been the middleware, or it may have been the client (i know swiftclient will delete segments by default when overwriting or deleting large objects. you can include a --leave-segments flag to skip that) | 20:42 |
| clayg | NM object-expirer? | 20:44 |
| clayg | ssync has this weird bug where it sync's expired object .data files as tombstones -> https://bugs.launchpad.net/swift/+bug/1652323 | 20:45 |
| openstack | Launchpad bug 1652323 in OpenStack Object Storage (swift) "ssync syncs an expired object as a tombstone, probe test_object_expirer fails" [Medium,Confirmed] | 20:45 |
| NM | timburke: Well, I couldn't find any delete request on the proxy. clayg: That is my first suspect but according to the logs, our client doesn't use expires. One thing that I'll check is if I can find any DELETE on the object services since I couldn't find on the proxies. | 20:52 |
| NM | timburke: clayg: thanks for the help. I need to get my kid. If I find anything interesting I'll share tomorrow. | 20:53 |
| *** NM has quit IRC | 20:53 | |
| *** eckesicle__ has joined #openstack-swift | 20:55 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Pretend *some* parts min_part_hours_passed https://review.openstack.org/311226 | 20:55 |
| *** sams-gle_ has joined #openstack-swift | 20:58 | |
| *** sams-gleb has quit IRC | 20:58 | |
| *** eckesicle__ is now known as eckesicle | 20:59 | |
| *** JimCheung has quit IRC | 21:00 | |
| *** JimCheung has joined #openstack-swift | 21:00 | |
| *** PavelK has quit IRC | 21:03 | |
| *** dfflanders has joined #openstack-swift | 21:03 | |
| *** JimCheung has quit IRC | 21:04 | |
| *** dmorita has quit IRC | 21:05 | |
| *** dmorita has joined #openstack-swift | 21:06 | |
| *** dmorita has quit IRC | 21:06 | |
| *** dmorita has joined #openstack-swift | 21:07 | |
| *** cdelatte has quit IRC | 21:17 | |
| *** clu_ has joined #openstack-swift | 21:22 | |
| *** JimCheung has joined #openstack-swift | 21:31 | |
| *** sams-gle_ has quit IRC | 21:34 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** dmorita has quit IRC | 21:37 | |
| *** dmorita has joined #openstack-swift | 21:37 | |
| *** dmorita_ has joined #openstack-swift | 21:38 | |
| *** dmorita has quit IRC | 21:38 | |
| *** Jeffrey4l has joined #openstack-swift | 21:48 | |
| mattoliverau | morning | 21:56 |
| *** d0ugal has quit IRC | 21:59 | |
| timburke | good morning mattoliverau! | 21:59 |
| timburke | i guess notmyname ought to show up soon :-) | 21:59 |
| *** d0ugal has joined #openstack-swift | 21:59 | |
| *** d0ugal has quit IRC | 21:59 | |
| *** d0ugal has joined #openstack-swift | 21:59 | |
| *** tqtran has joined #openstack-swift | 22:06 | |
| *** tqtran has quit IRC | 22:08 | |
| *** catintheroof has quit IRC | 22:20 | |
| *** vint_bra has quit IRC | 22:33 | |
| *** sams-gleb has joined #openstack-swift | 22:35 | |
| *** sams-gleb has quit IRC | 22:39 | |
| *** david-lyle has quit IRC | 22:51 | |
| *** andymccr has quit IRC | 23:20 | |
| *** asettle has quit IRC | 23:20 | |
| *** klamath has quit IRC | 23:27 | |
| *** kei_yama has joined #openstack-swift | 23:33 | |
| *** chsc has quit IRC | 23:36 | |
| *** lespaul has joined #openstack-swift | 23:42 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!