| *** klamath has quit IRC | 00:05 | |
| *** klamath has joined #openstack-swift | 00:08 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift master: Refactor object updater's async-pending finder https://review.openstack.org/571343 | 00:39 |
|---|---|---|
| *** gerardgine has quit IRC | 00:40 | |
| *** amandap_ has joined #openstack-swift | 00:58 | |
| *** amandap has quit IRC | 01:00 | |
| *** gyankum has joined #openstack-swift | 01:19 | |
| openstackgerrit | Merged openstack/swift master: Added PyPI validation check for the readme. https://review.openstack.org/571215 | 01:22 |
| *** amandap_ has quit IRC | 01:29 | |
| *** lifeless_ has joined #openstack-swift | 01:40 | |
| *** lifeless has quit IRC | 01:40 | |
| *** mvk has quit IRC | 02:06 | |
| *** mvk has joined #openstack-swift | 02:19 | |
| openstackgerrit | Merged openstack/swift master: removed specific reference to two core reviewers https://review.openstack.org/571328 | 02:32 |
| *** klamath has quit IRC | 02:39 | |
| *** klamath has joined #openstack-swift | 03:13 | |
| *** klamath has quit IRC | 03:29 | |
| *** amandap has joined #openstack-swift | 03:29 | |
| *** rcernin has quit IRC | 03:35 | |
| *** kei_yama has quit IRC | 03:43 | |
| *** ianychoi has quit IRC | 03:51 | |
| *** rcernin has joined #openstack-swift | 03:54 | |
| *** mvk has quit IRC | 03:56 | |
| *** kei_yama has joined #openstack-swift | 04:01 | |
| *** gyee has quit IRC | 04:21 | |
| *** rcernin has quit IRC | 04:28 | |
| *** rcernin has joined #openstack-swift | 04:29 | |
| *** links has joined #openstack-swift | 04:38 | |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: Make ShardRanges and its CLI py3-compliant https://review.openstack.org/570320 | 05:12 |
| *** cshastri has joined #openstack-swift | 05:16 | |
| *** cshastri has quit IRC | 05:28 | |
| *** pcaruana has joined #openstack-swift | 06:43 | |
| *** klamath has joined #openstack-swift | 06:44 | |
| kota_ | hello, perhaps acoles knows, is there any chance to run swift-container-sharder w/o internal-client.conf located under /etc/swift ? | 06:47 |
| kota_ | i noticed storlets gate broken that attempts to setup Swift via devstack but no configuration exists for internal-client then swift-init all start got an error like No such file or directory for container-sharder. | 06:49 |
| *** klamath has quit IRC | 06:49 | |
| kota_ | hmmm... anyway, fixing the devstack seems a right way tho. | 06:52 |
| *** rcernin has quit IRC | 07:11 | |
| kota_ | https://review.openstack.org/571391 | 07:11 |
| patchbot | patch 571391 - openstack-dev/devstack - Add internal-client config integration for Swift | 07:11 |
| *** tesseract has joined #openstack-swift | 07:22 | |
| *** ianychoi has joined #openstack-swift | 07:32 | |
| *** geaaru has joined #openstack-swift | 07:40 | |
| *** cbartz has joined #openstack-swift | 07:41 | |
| *** ccamacho has joined #openstack-swift | 07:44 | |
| *** amandap has quit IRC | 07:45 | |
| *** amandap has joined #openstack-swift | 07:45 | |
| *** amandap has quit IRC | 07:48 | |
| *** amandap has joined #openstack-swift | 07:48 | |
| acoles | kota_: the path to the internal client conf is configurable | 07:59 |
| acoles | kota_: https://github.com/openstack/swift/blob/26418140108281ae5ac11004ebf33b7b3e08c74d/etc/container-server.conf-sample#L409-L411 | 08:00 |
| kota_ | acoles: yes, it's configurable but if it doesn't exist, the sharder cannot start the work. | 08:01 |
| kota_ | acoles: but, thanks. | 08:01 |
| acoles | kota_: OIC. no, it is required for the sharder | 08:02 |
| kota_ | and it seems like that i did duplicate work with https://review.openstack.org/#/c/570869/ :P | 08:03 |
| patchbot | patch 570869 - openstack-dev/devstack - swift expects an internal client config to start | 08:03 |
| *** hoonetorg has quit IRC | 08:11 | |
| *** rcernin has joined #openstack-swift | 08:12 | |
| *** rcernin has quit IRC | 08:12 | |
| acoles | kota_: internal client is also required for container sync I think? | 08:12 |
| acoles | ah no, container sync has an inline default config | 08:13 |
| acoles | so the conf file is not *required* | 08:13 |
| kota_ | OIC, we have dsvm gate but it does never fail. the reason seems like, we have an option SWIFT_START_ALL_SERVICES=False for devstack | 08:13 |
| kota_ | http://logs.openstack.org/69/570869/2/check/legacy-swift-dsvm-functional/837bd78/logs/local.conf.txt.gz | 08:13 |
| kota_ | acoles: that's what I thought about container-sync | 08:14 |
| kota_ | I'm not sure if it's *fortunate* or *unfortunate* | 08:14 |
| *** mikecmpbll has joined #openstack-swift | 08:15 | |
| *** hoonetorg has joined #openstack-swift | 08:24 | |
| kota_ | timburke, notmyname: perhaps, p 570869 hits the eventlet 0.23.0 issue that breaks swift ??? | 08:27 |
| patchbot | https://review.openstack.org/#/c/570869/ - openstack-dev/devstack - swift expects an internal client config to start | 08:27 |
| kota_ | no | 08:27 |
| kota_ | no it's not the patch i meant | 08:27 |
| kota_ | i meant, p 564700 | 08:28 |
| patchbot | https://review.openstack.org/#/c/564700/ - swift - WIP: Add ceph-s3 test non-voting job | 08:28 |
| kota_ | I just saw notmyname's mail in the openstack-dev mailing list -> http://lists.openstack.org/pipermail/openstack-dev/2018-May/130256.html | 08:30 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: WIP: Add ceph-s3 test non-voting job https://review.openstack.org/564700 | 08:32 |
| *** lifeless_ is now known as lifeless | 08:42 | |
| *** klamath has joined #openstack-swift | 08:46 | |
| *** klamath has quit IRC | 08:51 | |
| *** mikecmpb_ has joined #openstack-swift | 09:08 | |
| *** mikecmpbll has quit IRC | 09:09 | |
| *** amandap has quit IRC | 09:10 | |
| remix_tj | notmyname: just to let you know, EMC implements retention header with a non standard extension: http://doc.isilon.com/ECS/3.2/DataAccessGuide/ecs_c_swift_reterntion.html so my idea is to make this a standard option :-) | 09:23 |
| *** silor has joined #openstack-swift | 09:35 | |
| *** lifeless has quit IRC | 09:55 | |
| *** lifeless has joined #openstack-swift | 09:56 | |
| *** mikecmpb_ has quit IRC | 09:58 | |
| *** mikecmpbll has joined #openstack-swift | 09:59 | |
| *** silor1 has joined #openstack-swift | 10:00 | |
| *** silor has quit IRC | 10:03 | |
| *** silor1 is now known as silor | 10:03 | |
| *** lifeless has quit IRC | 10:08 | |
| *** lifeless_ has joined #openstack-swift | 10:08 | |
| *** silor has quit IRC | 10:10 | |
| *** armaan has joined #openstack-swift | 10:33 | |
| *** klamath has joined #openstack-swift | 10:47 | |
| *** klamath has quit IRC | 10:51 | |
| *** amandap has joined #openstack-swift | 10:52 | |
| *** gyankum has quit IRC | 11:00 | |
| *** tesseract has quit IRC | 11:07 | |
| *** tesseract has joined #openstack-swift | 11:08 | |
| *** sajjadg has joined #openstack-swift | 11:25 | |
| *** mikecmpbll has quit IRC | 11:47 | |
| *** mikecmpbll has joined #openstack-swift | 11:49 | |
| *** armaan has quit IRC | 11:50 | |
| *** armaan has joined #openstack-swift | 11:50 | |
| *** armaan has quit IRC | 12:13 | |
| *** armaan has joined #openstack-swift | 12:14 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Fix marker and end_marker descriptions in API ref https://review.openstack.org/571460 | 12:29 |
| *** kei_yama has quit IRC | 12:29 | |
| *** sajjadg has quit IRC | 12:39 | |
| *** prasen has joined #openstack-swift | 12:41 | |
| *** prasen has quit IRC | 12:48 | |
| *** prasen has joined #openstack-swift | 12:49 | |
| *** sajjadg has joined #openstack-swift | 12:50 | |
| *** zaitcev has quit IRC | 12:57 | |
| *** jistr is now known as jistr|mtg | 12:59 | |
| *** amandap has quit IRC | 12:59 | |
| *** armaan has quit IRC | 13:11 | |
| *** armaan has joined #openstack-swift | 13:12 | |
| *** armaan has quit IRC | 13:17 | |
| *** zaitcev has joined #openstack-swift | 13:17 | |
| *** ChanServ sets mode: +v zaitcev | 13:17 | |
| *** armaan has joined #openstack-swift | 13:23 | |
| *** links has quit IRC | 13:29 | |
| *** amandap has joined #openstack-swift | 13:32 | |
| *** klamath has joined #openstack-swift | 13:36 | |
| *** armaan has quit IRC | 13:38 | |
| *** gyankum has joined #openstack-swift | 13:42 | |
| *** mvenesio has joined #openstack-swift | 13:43 | |
| openstackgerrit | Merged openstack/swift master: Stop overriding _in_sync in container replicator https://review.openstack.org/570055 | 14:05 |
| *** olc has quit IRC | 14:08 | |
| *** mvenesio has quit IRC | 14:13 | |
| *** amandap has quit IRC | 14:14 | |
| *** prasen has quit IRC | 14:27 | |
| *** armaan has joined #openstack-swift | 14:46 | |
| *** jistr|mtg is now known as jistr | 14:57 | |
| *** sajjadg has quit IRC | 15:09 | |
| *** silor has joined #openstack-swift | 15:10 | |
| *** silor has quit IRC | 15:13 | |
| *** ccamacho has quit IRC | 15:15 | |
| *** ccamacho has joined #openstack-swift | 15:15 | |
| *** itlinux has joined #openstack-swift | 15:25 | |
| *** gyee has joined #openstack-swift | 15:26 | |
| *** pcaruana has quit IRC | 15:28 | |
| *** itlinux has quit IRC | 15:31 | |
| *** itlinux has joined #openstack-swift | 15:31 | |
| *** mvk has joined #openstack-swift | 15:54 | |
| *** amandap has joined #openstack-swift | 16:03 | |
| *** itlinux has quit IRC | 16:07 | |
| *** cbartz has quit IRC | 16:11 | |
| *** hseipp has joined #openstack-swift | 16:16 | |
| *** ccamacho has quit IRC | 16:21 | |
| *** hseipp has quit IRC | 16:21 | |
| notmyname | good morning | 16:23 |
| *** itlinux has joined #openstack-swift | 16:24 | |
| *** geaaru has quit IRC | 16:27 | |
| *** hseipp has joined #openstack-swift | 16:27 | |
| timburke | good morning | 16:35 |
| timburke | fwiw, the default internal client config for container sync was done for backwards compatibility, so you could upgrade swift before updating config and container sync would still work | 16:36 |
| timburke | the container sharder, being a new daemon, doesn't have to worry about the upgrade path | 16:38 |
| *** mikecmpbll has quit IRC | 16:40 | |
| *** armaan has quit IRC | 16:47 | |
| *** hseipp has quit IRC | 16:47 | |
| *** armaan has joined #openstack-swift | 16:48 | |
| timburke | remix_tj: interesting... i've played around with something like setting a retention period for a container -- one of the use-cases that drove me to write https://review.openstack.org/#/c/342857/ was being able to set a default x-delete-after header on PUT to a particular container | 16:48 |
| patchbot | patch 342857 - swift - Add defaulter middleware | 16:48 |
| *** amandap has quit IRC | 16:57 | |
| *** prasen has joined #openstack-swift | 17:02 | |
| timburke | that layer of indirection is interesting, though -- being able to change the retention period for *a group* objects by changing the value for the policy is kinda cool | 17:09 |
| *** klamath_ has joined #openstack-swift | 17:11 | |
| *** klamath__ has joined #openstack-swift | 17:12 | |
| notmyname | timburke: can you relink what remix_tj was referring to? my IRC client closed this morning before I could read my znc buffer | 17:13 |
| timburke | notmyname: http://doc.isilon.com/ECS/3.2/DataAccessGuide/ecs_c_swift_reterntion.html | 17:14 |
| notmyname | thanks | 17:14 |
| *** klamath has quit IRC | 17:14 | |
| *** klamath_ has quit IRC | 17:15 | |
| notmyname | along those lines, I've been pondering clayg's idea of having a different `unlink_older_than()`. ie an `unlink_newer_than()`, and use that for some sort of "retention" | 17:16 |
| notmyname | it's an interesting compromise for eventually-consistent storage that has a "no overwrites or deletes" policy/rule | 17:16 |
| notmyname | turns into "well maybe you can overwrite data, but we'll get it sorted and give you back the old stuff eventually" | 17:17 |
| notmyname | sounds simpler (and more "swift-y") than adding an etcd layer to the system to get paxos/raft | 17:18 |
| *** klamath__ has quit IRC | 17:18 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Add detail to container server GET docstring https://review.openstack.org/571530 | 17:19 |
| openstackgerrit | Alistair Coles proposed openstack/swift master: WIP Always fetch listings from shards when possible https://review.openstack.org/571531 | 17:19 |
| *** klamath has joined #openstack-swift | 17:22 | |
| timburke | notmyname: i remember you (or maybe torgomatic?) mentioning once that we could probably come up with a way to have a storage policy for WORM that basically runs the consistency bits in reverse... i still rather like that idea | 17:22 |
| notmyname | I don't think it was me. I thought it was clayg. could have been torgomatic. yeah, it's a pretty cool idea | 17:23 |
| *** klamath_ has joined #openstack-swift | 17:24 | |
| tdasilva | isn't the issue with that idea that you give a user a 200 response for something that won't be there eventually | 17:27 |
| tdasilva | ? | 17:27 |
| *** klamath has quit IRC | 17:28 | |
| *** klamath has joined #openstack-swift | 17:29 | |
| *** klamath_ has quit IRC | 17:31 | |
| notmyname | I don't know how much of an issue that is. I mean, what happens if you send a PUT today, but 15 ms before another PUT starts. you'll get a 200, but your data won't be there | 17:31 |
| *** klamath_ has joined #openstack-swift | 17:32 | |
| notmyname | isn't it a question of how it responds when there are failures? and in that case today, there's valid debate over if what we do is "right" or not | 17:32 |
| notmyname | I think simply writing down what happens and then sticking to it may be sufficient | 17:33 |
| timburke | (obvs, of course, it should respond 409 if there *aren't* failures) | 17:33 |
| notmyname | tdasilva: I mean, that's a completely valid point and a really really good thing we should discuss | 17:33 |
| notmyname | timburke: obviously ;-) | 17:33 |
| *** klamath has quit IRC | 17:35 | |
| clayg | good morning | 17:35 |
| *** gyankum has quit IRC | 17:35 | |
| torgomatic | if you're going to do WORM and not lie to the user with 200s for things that vanish, you've got to have an immediately-consistent set somewhere that contains the names of all the objects which are immutably-stored. | 17:38 |
| timburke | torgomatic: but we *already* lie to the user about data that might vanish. we just hand-wave and say "eventual consistency" | 17:40 |
| timburke | why can't we do the same here? | 17:40 |
| *** d0ugal_ has joined #openstack-swift | 17:40 | |
| torgomatic | we could, or we could try to do better than our past selves :) | 17:40 |
| tdasilva | one way to do it is to rely on strong-consistent backends, that's probably what EMC is doing with isilon... | 17:42 |
| *** d0ugal has quit IRC | 17:42 | |
| torgomatic | if you have strong consistency *somewhere*, it's easy to do stuff like WORM | 17:43 |
| torgomatic | if all you have is eventual consistency everywhere, it's really tough | 17:43 |
| torgomatic | what if the user makes a WORM-policy container during a network failure, then PUTs some objects, but then it turns out that the WORM policy loses during container replication, so those "immutable" objects aren't? | 17:45 |
| *** d0ugal__ has joined #openstack-swift | 17:51 | |
| *** klamath_ has quit IRC | 17:51 | |
| *** klamath has joined #openstack-swift | 17:51 | |
| torgomatic | It probably depends on what you need from WORM. If you want pretty-good guarantees, then a policy-based approach is great. If you've got legal requirements that these documents MUST NOT get overwritten no matter what, then you're going to want the guarantee. | 17:53 |
| *** d0ugal_ has quit IRC | 17:53 | |
| *** itlinux has quit IRC | 17:53 | |
| notmyname | I laugh (on the inside) every time I hear someone give ultimatums about what MUST and MUST NOT happen in distributed systems | 17:54 |
| *** itlinux has joined #openstack-swift | 17:57 | |
| *** itlinux has quit IRC | 17:58 | |
| *** d0ugal__ has quit IRC | 17:59 | |
| torgomatic | the tools exist to make this particular thing happen with strong guarantees; they're just expensive and on the other side of CAP from where Swift is | 17:59 |
| notmyname | yeah | 18:00 |
| torgomatic | if I had an internal customer asking for WORM, I'd need to find out a bit more about their requirements before choosing the cheap first-write-wins in replication or the expensive index-of-all-objects option | 18:01 |
| *** itlinux has joined #openstack-swift | 18:05 | |
| *** armaan has quit IRC | 18:21 | |
| *** armaan has joined #openstack-swift | 18:21 | |
| *** d0ugal__ has joined #openstack-swift | 18:22 | |
| *** d0ugal__ has quit IRC | 18:34 | |
| *** tesseract has quit IRC | 18:34 | |
| *** lifeless_ has quit IRC | 18:39 | |
| *** ccamacho has joined #openstack-swift | 18:45 | |
| *** mvk has quit IRC | 18:52 | |
| *** amandap has joined #openstack-swift | 18:58 | |
| *** d0ugal__ has joined #openstack-swift | 18:59 | |
| *** amandap has quit IRC | 19:02 | |
| *** armaan has quit IRC | 19:04 | |
| timburke | i'm not *opposed* to taking that kind of an approach -- i think there are a lot of benefits to be had that would resolve a lot of the hard-to-paper-over differences with s3 -- but it seems like it may be beyond the scope of swift | 19:12 |
| *** prasen has quit IRC | 19:18 | |
| *** armaan has joined #openstack-swift | 19:37 | |
| *** lifeless has joined #openstack-swift | 19:38 | |
| *** armaan has quit IRC | 19:43 | |
| openstackgerrit | John Dickinson proposed openstack/swift master: Added S3 compatibility information to docs https://review.openstack.org/571561 | 19:51 |
| *** armaan has joined #openstack-swift | 19:56 | |
| *** armaan has quit IRC | 20:03 | |
| openstackgerrit | John Dickinson proposed openstack/swift master: Added S3 compatibility information to docs https://review.openstack.org/571561 | 20:19 |
| *** lifeless_ has joined #openstack-swift | 20:36 | |
| *** lifeless has quit IRC | 20:36 | |
| *** mikecmpbll has joined #openstack-swift | 21:06 | |
| *** links has joined #openstack-swift | 21:42 | |
| *** hoonetorg has quit IRC | 21:55 | |
| *** itlinux has quit IRC | 22:04 | |
| *** hoonetorg has joined #openstack-swift | 22:08 | |
| *** rcernin has joined #openstack-swift | 22:13 | |
| *** links has quit IRC | 22:16 | |
| *** lifeless_ has quit IRC | 22:33 | |
| *** lifeless has joined #openstack-swift | 22:34 | |
| *** mikecmpbll has quit IRC | 22:40 | |
| *** d0ugal__ has quit IRC | 22:48 | |
| *** d0ugal__ has joined #openstack-swift | 23:02 | |
| *** mikecmpbll has joined #openstack-swift | 23:03 | |
| *** kei_yama has joined #openstack-swift | 23:09 | |
| *** amandap has joined #openstack-swift | 23:28 | |
| notmyname | tdasilva: FWIW https://github.com/notmyname/alpine_swift <-- what I've been picking at recently and what I want to include in swift | 23:34 |
| openstackgerrit | Merged openstack/swift master: Add detail to container server GET docstring https://review.openstack.org/571530 | 23:45 |
| openstackgerrit | Merged openstack/swift master: Remove some Python 2.6 compatibility code in ring https://review.openstack.org/570421 | 23:55 |
| openstackgerrit | Merged openstack/swift master: Use X-Timestamp when checking object expiration https://review.openstack.org/532595 | 23:55 |
| *** mikecmpbll has quit IRC | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!