| tdasilva | JimCheung: FWIW IMO a good rule of thumb is to try to break down bit changes into smaller patches if at all possible. IMO it makes it easier to review | 00:03 |
|---|---|---|
| tdasilva | s/bit/big | 00:03 |
| *** abhitechie has joined #openstack-swift | 00:04 | |
| JimCheung | Make sense. I know it'll tough to do a huge review. Will try to break it down to logical chunks. | 00:05 |
| *** abhinavtechie has quit IRC | 00:06 | |
| *** vint_bra has joined #openstack-swift | 00:06 | |
| *** tsg has quit IRC | 00:07 | |
| *** garyj has joined #openstack-swift | 00:08 | |
| *** asettle has joined #openstack-swift | 00:10 | |
| *** tsg has joined #openstack-swift | 00:12 | |
| *** asettle has quit IRC | 00:14 | |
| *** rcernin has quit IRC | 00:22 | |
| *** diogogmt has quit IRC | 00:35 | |
| *** tsg has quit IRC | 00:41 | |
| *** abhitechie has quit IRC | 00:57 | |
| *** mingyu has joined #openstack-swift | 01:09 | |
| *** asettle has joined #openstack-swift | 01:13 | |
| *** vint_bra has quit IRC | 01:16 | |
| *** asettle has quit IRC | 01:18 | |
| *** garyj has quit IRC | 01:20 | |
| *** garyj has joined #openstack-swift | 01:20 | |
| *** garyj has quit IRC | 01:24 | |
| *** garyj has joined #openstack-swift | 01:26 | |
| *** garyj has quit IRC | 01:26 | |
| kota_ | good morning | 01:34 |
| *** mingyu has quit IRC | 01:34 | |
| *** diogogmt has joined #openstack-swift | 01:36 | |
| notmyname | if someone's on twitter and know's the answer, please reply https://twitter.com/tjdett/status/808832476093087744 | 01:37 |
| notmyname | (to be fair, i'm not entirely clear on what he's asking) | 01:38 |
| tdasilva | notmyname: I wonder if that's timss from earlier today | 01:39 |
| tdasilva | notmyname: see scrollback | 01:39 |
| tdasilva | kota_: o/ | 01:44 |
| mattoliverau | kota_: morning | 01:44 |
| kota_ | mattoliverau, tdasilva: o/ | 01:48 |
| *** zhugaoxiao has quit IRC | 01:51 | |
| *** zhugaoxiao has joined #openstack-swift | 01:52 | |
| notmyname | tdasilva: oh interesting | 01:54 |
| notmyname | mattoliverau: looks like that guy on twitter is from your area (Uni queensland) | 01:56 |
| *** Jeffrey4l has quit IRC | 02:05 | |
| *** tqtran has quit IRC | 02:07 | |
| *** Jeffrey4l has joined #openstack-swift | 02:15 | |
| charz | kota_: mattoliverau tdasilva morning | 02:20 |
| kota_ | hi charz! | 02:20 |
| *** JimCheung has quit IRC | 02:37 | |
| *** mingyu has joined #openstack-swift | 02:38 | |
| mattoliverau | notmyname: I'll take a look. Sorry been lunching :) | 02:40 |
| mattoliverau | charz: o/ | 02:40 |
| *** mingyu has quit IRC | 02:48 | |
| *** mingyu has joined #openstack-swift | 02:59 | |
| *** dmorita has quit IRC | 03:00 | |
| *** dmorita has joined #openstack-swift | 03:03 | |
| *** _david_cole_soho has quit IRC | 03:07 | |
| *** dmorita has quit IRC | 03:07 | |
| *** kei_yama has quit IRC | 03:09 | |
| *** siva_krish has joined #openstack-swift | 03:19 | |
| *** dmorita has joined #openstack-swift | 03:20 | |
| *** dmorita has quit IRC | 03:25 | |
| *** gmmaha has quit IRC | 03:31 | |
| *** gmmaha has joined #openstack-swift | 03:31 | |
| *** _david_cole_soho has joined #openstack-swift | 03:48 | |
| *** si1ver has quit IRC | 03:53 | |
| *** tqtran has joined #openstack-swift | 04:03 | |
| *** dmorita has joined #openstack-swift | 04:05 | |
| jrichli | notmyname, kota: I just found out that Daniel Axtens of IBM has been working on optimizations for EC. His goal to provide a Power-optimised erasure coding solution that does not require a hardware accelerator. He has a set of patches for liberasurecode which streamline a lot of overhead, which speeds up EC - especially of small objects. He says there would be benefits outside of just Power. | 04:05 |
| jrichli | He got in touch with me to find out what the process is to submit these patches. I pointed him to our internal process and the recent IRC convo with JimCheung. He said the patches aren't 100% ready just yet, but he wanted to get started with the process necessary to submit. | 04:06 |
| *** gmmaha has quit IRC | 04:07 | |
| *** tqtran has quit IRC | 04:08 | |
| *** gmmaha has joined #openstack-swift | 04:09 | |
| *** klrmn1 has quit IRC | 04:15 | |
| *** psachin has joined #openstack-swift | 04:18 | |
| *** mingyu has quit IRC | 04:20 | |
| *** mingyu has joined #openstack-swift | 04:21 | |
| *** mingyu has quit IRC | 04:26 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift: Split up tests for Concurrent Gets https://review.openstack.org/286589 | 04:32 |
| mattoliverau | clayg: ^^ I had a play with this.. and then got lost down the rabbit hole. | 04:32 |
| mattoliverau | jrichli: cool! | 04:33 |
| *** JimCheung has joined #openstack-swift | 04:37 | |
| *** JimCheung has quit IRC | 04:42 | |
| *** chlong has joined #openstack-swift | 04:44 | |
| jrichli | mattoliverau: he is an Aussie :-) | 04:44 |
| mattoliverau | jrichli: nice | 04:45 |
| mattoliverau | jrichli: and by power do you mean power arch? or power. Because IBM + Australian + Power arch, usually means IBM OzLabs :) | 04:46 |
| jrichli | mattoliverau: I think power as in open power. Does OzLabs do open power? | 04:48 |
| jrichli | I am not sure if he works locally | 04:48 |
| *** ppai has joined #openstack-swift | 04:48 | |
| jrichli | OzLabs sounds cool | 04:49 |
| kota_ | jrichli: I'm not faimilier with OpenPower but it sounds like it is related to something like SIMD accelaration? | 04:49 |
| jrichli | kota_: you know more than I do! I know of a brownbag that was presented at Barcelona for a swift cluster on open power. He works with those people a bit, but I get the impression not on that direct team. | 04:52 |
| *** JimCheung has joined #openstack-swift | 04:53 | |
| kota_ | jrichli: i saw the video, wait a bit finding the link | 04:54 |
| kota_ | jrichli: i found mattoliverau before finding the video link https://www.openstack.org/assets/summits/highlights/pics/30764570252-b7ca2c3d3d-z.jpg | 04:54 |
| jrichli | kota_: nice! | 04:55 |
| jrichli | kota_: https://www.youtube.com/watch?v=g2hO-YLXqGk | 04:56 |
| kota_ | jrichli: oh yeah, that one. thx | 04:56 |
| jrichli | but again, I am not sure how much he is involved in that cluster specifically | 04:57 |
| *** JimCheung has quit IRC | 04:57 | |
| *** links has joined #openstack-swift | 04:57 | |
| kota_ | jrichli: ok | 04:58 |
| jrichli | kota_: there also was https://www.youtube.com/watch?v=Vg0ujfLpIN8 | 04:58 |
| kota_ | thx! | 05:00 |
| jrichli | well, I am going to sleep now. have a good day | 05:00 |
| kota_ | good night jrichli! | 05:00 |
| mattoliverau | jrichli: yeah they do alot of open power :) | 05:05 |
| *** tqtran has joined #openstack-swift | 05:06 | |
| mattoliverau | kota_: \o/ its me | 05:06 |
| mattoliverau | tonyb: ^^ more openpower stuff! | 05:07 |
| notmyname | cool | 05:07 |
| mattoliverau | jrichli: tonyb used to work at IBM on open power, he's one of the openpower kernel devs (or was one I guess). | 05:07 |
| tonyb | mattoliverau: yeah past life. Now I just randomly hack on OpenStack to make it work on openpower ;P | 05:08 |
| *** dmorita has quit IRC | 05:09 | |
| *** tqtran has quit IRC | 05:10 | |
| mattoliverau | tonyb: which is awesome :) | 05:12 |
| tonyb | :) | 05:12 |
| *** asettle has joined #openstack-swift | 05:14 | |
| *** douglascorrea has quit IRC | 05:14 | |
| *** dmorita has joined #openstack-swift | 05:14 | |
| *** douglascorrea has joined #openstack-swift | 05:15 | |
| *** garyj has joined #openstack-swift | 05:15 | |
| *** douglascorrea has quit IRC | 05:15 | |
| *** douglascorrea has joined #openstack-swift | 05:16 | |
| tonyb | mattoliverau: https://youtu.be/Vg0ujfLpIN8?t=7m37s Openpower and Rackspace :) | 05:16 |
| *** asettle has quit IRC | 05:21 | |
| *** dmorita has quit IRC | 05:26 | |
| *** JimCheung has joined #openstack-swift | 05:33 | |
| *** JimCheung has quit IRC | 05:37 | |
| *** dja has joined #openstack-swift | 05:44 | |
| *** garyj has quit IRC | 05:48 | |
| *** garyj has joined #openstack-swift | 05:48 | |
| *** diogogmt has quit IRC | 05:58 | |
| *** diogogmt has joined #openstack-swift | 05:58 | |
| *** garyj has quit IRC | 06:17 | |
| *** mingyu has joined #openstack-swift | 06:18 | |
| *** McMurlock1 has joined #openstack-swift | 06:31 | |
| *** ChubYann has quit IRC | 07:01 | |
| *** douglascorrea has quit IRC | 07:02 | |
| *** douglascorrea has joined #openstack-swift | 07:03 | |
| *** maestropandy has joined #openstack-swift | 07:07 | |
| *** douglascorrea has quit IRC | 07:07 | |
| *** McMurlock1 has quit IRC | 07:08 | |
| *** McMurlock1 has joined #openstack-swift | 07:08 | |
| *** sams-gleb has joined #openstack-swift | 07:08 | |
| *** siva_krish has quit IRC | 07:11 | |
| openstackgerrit | Tone Zhang proposed openstack/swift: Update the tag for Swift tox test https://review.openstack.org/410560 | 07:13 |
| *** McMurlock1 has quit IRC | 07:13 | |
| *** maestropandy has left #openstack-swift | 07:14 | |
| *** siva_krish has joined #openstack-swift | 07:15 | |
| openstackgerrit | Tone Zhang proposed openstack/swift: Update the tag for Swift tox test https://review.openstack.org/410560 | 07:18 |
| *** asettle has joined #openstack-swift | 07:18 | |
| *** bapalm has quit IRC | 07:18 | |
| *** asettle has quit IRC | 07:25 | |
| *** dmorita has joined #openstack-swift | 07:27 | |
| *** dmorita has quit IRC | 07:31 | |
| *** mingyu has quit IRC | 07:37 | |
| *** bapalm has joined #openstack-swift | 07:38 | |
| zaitcev | http://www.itnews.com.au/news/hpe-storage-crash-killed-ato-online-services-444490 - When I saw that HPE was the vendor that lost a petabyte, I was concerned that Swift was somehow involved. But apprently it's just 3PAR.... GJ tho lol | 07:39 |
| zaitcev | "iTnews understands the data corruption issues from the primary SAN were replicated onto the second SAN" | 07:39 |
| *** mingyu has joined #openstack-swift | 07:45 | |
| *** garyj has joined #openstack-swift | 07:50 | |
| *** mingyu_ has joined #openstack-swift | 07:52 | |
| *** mingyu has quit IRC | 07:54 | |
| *** mingyu_ has quit IRC | 07:55 | |
| *** rcernin has joined #openstack-swift | 07:56 | |
| *** mingyu has joined #openstack-swift | 07:56 | |
| *** rcernin has quit IRC | 07:56 | |
| *** rcernin has joined #openstack-swift | 07:57 | |
| *** pcaruana has joined #openstack-swift | 08:00 | |
| *** garyj has quit IRC | 08:01 | |
| *** tqtran has joined #openstack-swift | 08:08 | |
| *** tqtran has quit IRC | 08:12 | |
| *** sams-gleb has quit IRC | 08:17 | |
| *** sams-gleb has joined #openstack-swift | 08:18 | |
| *** sams-gleb has quit IRC | 08:22 | |
| *** oshritf has joined #openstack-swift | 08:26 | |
| *** JimCheung has joined #openstack-swift | 08:26 | |
| *** JimCheung has quit IRC | 08:30 | |
| *** daemontool has joined #openstack-swift | 08:44 | |
| *** rledisez has joined #openstack-swift | 08:44 | |
| *** openstackgerrit has quit IRC | 08:48 | |
| *** mingyu has quit IRC | 08:49 | |
| *** _david_cole_soho has quit IRC | 08:49 | |
| *** _david_cole_soho has joined #openstack-swift | 08:49 | |
| *** mingyu has joined #openstack-swift | 08:50 | |
| *** geaaru has joined #openstack-swift | 08:50 | |
| *** daemontool_ has joined #openstack-swift | 08:55 | |
| *** amoralej|off is now known as amoralej | 08:56 | |
| *** daemontool has quit IRC | 08:58 | |
| *** garyj has joined #openstack-swift | 09:02 | |
| *** mingyu has quit IRC | 09:04 | |
| *** mingyu has joined #openstack-swift | 09:04 | |
| *** joeljwright has joined #openstack-swift | 09:05 | |
| *** ChanServ sets mode: +v joeljwright | 09:05 | |
| *** garyj has quit IRC | 09:06 | |
| *** jordanP has joined #openstack-swift | 09:18 | |
| *** furlongm_ has joined #openstack-swift | 09:29 | |
| *** furlongm has quit IRC | 09:30 | |
| *** mingyu has quit IRC | 09:34 | |
| *** mingyu has joined #openstack-swift | 09:36 | |
| *** asettle has joined #openstack-swift | 09:44 | |
| *** furlongm has joined #openstack-swift | 09:54 | |
| *** cbartz has joined #openstack-swift | 09:55 | |
| *** furlongm_ has quit IRC | 09:55 | |
| *** dja has quit IRC | 09:56 | |
| *** sams-gleb has joined #openstack-swift | 10:01 | |
| *** mingyu has quit IRC | 10:02 | |
| *** siva_krish has quit IRC | 10:05 | |
| *** tqtran has joined #openstack-swift | 10:09 | |
| *** siva_krish has joined #openstack-swift | 10:10 | |
| *** tqtran has quit IRC | 10:13 | |
| *** dfflanders has quit IRC | 10:19 | |
| *** mvk has quit IRC | 10:24 | |
| *** links has quit IRC | 10:38 | |
| *** links has joined #openstack-swift | 10:38 | |
| *** mvk has joined #openstack-swift | 10:56 | |
| *** garyj has joined #openstack-swift | 11:03 | |
| *** garyj has quit IRC | 11:08 | |
| *** acoles_ is now known as acoles | 11:14 | |
| *** furlongm_ has joined #openstack-swift | 11:41 | |
| *** furlongm has quit IRC | 11:41 | |
| acoles | _david_cole_soho: about your invalid etags...you should also consider that any object metadata (x-object-meta-*) that may have been PUT with the object will not have been persisted by the older object server because encryption maps it to x-object-transient-sysmeta namespace, and older object servers don't know about that. | 11:50 |
| acoles | _david_cole_soho: so IF you had any such metadata (e.g. swiftclient adds file mtime as metadata on a PUT) then one object replica may not have it persisted, and fixing up etags in container db rows may not be enough. A get/put of a good replica would be safest. | 11:54 |
| acoles | hmmm, getting a *good replica* is significant, so any get/put script should check for expected metadata and retry if it's not included with the get response | 11:56 |
| *** tqtran has joined #openstack-swift | 12:11 | |
| *** tqtran has quit IRC | 12:15 | |
| *** openstackgerrit has joined #openstack-swift | 12:18 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Make tempurl functional tests clean up account keys https://review.openstack.org/346752 | 12:18 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Remove duplicate setup code from test_tempurl.py https://review.openstack.org/410731 | 12:42 |
| *** jamielennox is now known as jamielennox|away | 12:43 | |
| *** jamielennox|away is now known as jamielennox | 12:50 | |
| *** vint_bra has joined #openstack-swift | 12:51 | |
| *** amoralej is now known as amoralej|lunch | 12:55 | |
| *** garyj has joined #openstack-swift | 13:05 | |
| *** mathiasb has quit IRC | 13:06 | |
| *** garyj has quit IRC | 13:09 | |
| *** mathiasb has joined #openstack-swift | 13:13 | |
| *** mvk has quit IRC | 13:15 | |
| *** daemontool_ has quit IRC | 13:30 | |
| *** chlong has quit IRC | 13:32 | |
| *** klamath has joined #openstack-swift | 13:33 | |
| *** klamath has quit IRC | 13:33 | |
| *** klamath has joined #openstack-swift | 13:34 | |
| *** ppai has quit IRC | 13:38 | |
| *** tongli has joined #openstack-swift | 13:56 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Confirm receipt of SLO PUT with etag https://review.openstack.org/390901 | 14:01 |
| *** amoralej|lunch is now known as amoralej | 14:07 | |
| *** klamath has quit IRC | 14:11 | |
| *** tqtran has joined #openstack-swift | 14:12 | |
| *** klamath has joined #openstack-swift | 14:17 | |
| *** tqtran has quit IRC | 14:17 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Move documented reclaim_age option to correct location https://review.openstack.org/374419 | 14:19 |
| *** dmorita has joined #openstack-swift | 14:27 | |
| *** asettle has quit IRC | 14:31 | |
| *** dmorita has quit IRC | 14:32 | |
| *** asettle has joined #openstack-swift | 14:32 | |
| mathiasb | what's the difference between x-storage-token and x-auth-token as used in swift? | 14:34 |
| tdasilva | i don't think there's a difference, right? | 14:36 |
| tdasilva | doc says: "The token can be passed into Swift using the X-Auth-Token or the X-Storage-Token header" | 14:36 |
| *** mvk has joined #openstack-swift | 14:36 | |
| mathiasb | in all the examples I see, the values are the same | 14:36 |
| mathiasb | is there e.g., a difference that in some deployments one is required and in another, the other (e.g., depending on the authentication service)? | 14:37 |
| mathiasb | or can you always just pass one or the other and everything should behave the same? | 14:38 |
| *** JimCheung has joined #openstack-swift | 14:38 | |
| tdasilva | not that i know of, I imagine the reason for having both must have been an API thing where one was added first and then somebody asked to add the other to conform with some openstack api (maybe keystone) | 14:40 |
| tdasilva | like it was recently done with the transaction id header | 14:40 |
| mathiasb | ah ok, that makes sense | 14:41 |
| mathiasb | thanks! | 14:41 |
| tdasilva | notmyname might have the history better on this | 14:41 |
| *** JimCheung has quit IRC | 14:43 | |
| *** sams-gleb has quit IRC | 14:44 | |
| *** sams-gleb has joined #openstack-swift | 14:45 | |
| *** ChanServ sets mode: +v tdasilva | 14:48 | |
| *** tongli has quit IRC | 14:49 | |
| *** sams-gleb has quit IRC | 14:49 | |
| *** links has quit IRC | 14:57 | |
| *** mvk has quit IRC | 15:02 | |
| *** sams-gleb has joined #openstack-swift | 15:04 | |
| *** garyj has joined #openstack-swift | 15:07 | |
| *** psachin has quit IRC | 15:11 | |
| *** garyj has quit IRC | 15:11 | |
| *** mvk has joined #openstack-swift | 15:16 | |
| *** tsg has joined #openstack-swift | 15:19 | |
| *** chlong has joined #openstack-swift | 15:21 | |
| *** cebreidian_ has quit IRC | 15:30 | |
| *** siva_krish has quit IRC | 15:39 | |
| *** siva_krish has joined #openstack-swift | 15:40 | |
| *** _david_cole_soho has quit IRC | 15:41 | |
| *** eckesicle__ has joined #openstack-swift | 15:44 | |
| eckesicle__ | Hi everyone | 15:45 |
| eckesicle__ | i'm seeing an interesting issue that appears to be affecting only one user in my swift cluster | 15:45 |
| eckesicle__ | basically | 15:45 |
| eckesicle__ | Dec 14 08:31:28 sld-stor2 proxy-server: 10.11.1.103 10.11.1.103 14/Dec/2016/08/31/28 GET /v1/AUTH_fe48f5d760a14d1a9c398d95275235bf/67554E9F71175344DE11CFAF/heartbeat HTTP/1.0 404 - fog/1.38.0%20fog-core/1.40.0 691ea90725bc4d2f... - 70 - txe47d85c304bf435c9a7a8-0058510360 - 0.2220 - - 1481704288.135571957 1481704288.357614994 0 | 15:45 |
| eckesicle__ | this is the log entry im seeing | 15:45 |
| eckesicle__ | all files (migrated from an old cluster) | 15:46 |
| eckesicle__ | for a particular user is returning 404s | 15:46 |
| eckesicle__ | BUT | 15:46 |
| eckesicle__ | if that user uploads a file to the container | 15:46 |
| eckesicle__ | that file is then accessible | 15:46 |
| *** siva_krish has quit IRC | 15:48 | |
| eckesicle__ | it's not an auth issue, because the admin is unable to see the files too | 15:48 |
| eckesicle__ | even though uploading those files were definitely successful | 15:49 |
| acoles | eckesicle__: how were the files migrated? | 15:52 |
| *** dja has joined #openstack-swift | 15:52 | |
| eckesicle__ | They were migrated using a ruby script | 15:52 |
| eckesicle__ | sec | 15:53 |
| *** diogogmt has quit IRC | 15:53 | |
| eckesicle__ | https://gist.github.com/cfeckardt/addf267196d7173fa5037e3f7ddd0b83 | 15:53 |
| eckesicle__ | (it using the Openstack fog wrapper | 15:54 |
| eckesicle__ | (it uses* the Openstack fog wrapper) | 15:54 |
| *** dja has quit IRC | 15:57 | |
| *** _david_cole_soho has joined #openstack-swift | 15:58 | |
| *** tongli has joined #openstack-swift | 16:00 | |
| acoles | eckesicle__: the swift-get-nodes command will tell you where the object files should be, so you can verify the files exist | 16:01 |
| eckesicle__ | thanks | 16:01 |
| eckesicle__ | ill have a look | 16:01 |
| *** cbartz has quit IRC | 16:05 | |
| acoles | also check that the objects are shown in the container listing. if the container db didn't get migrated then the GETs might be using the wrong (default) storage policy | 16:08 |
| *** siva_krish has joined #openstack-swift | 16:10 | |
| eckesicle__ | thanks acoles | 16:10 |
| eckesicle__ | but i didnt migrate the containers | 16:10 |
| eckesicle__ | i created new containers with new ids | 16:10 |
| eckesicle__ | im going to do a little more digging with this | 16:11 |
| acoles | swift-object-info is another useful tool to verify object files | 16:11 |
| *** mingyu has joined #openstack-swift | 16:12 | |
| *** rcernin has quit IRC | 16:14 | |
| *** _JZ_ has joined #openstack-swift | 16:15 | |
| eckesicle__ | acoles so im actually unable to find the container | 16:22 |
| eckesicle__ | swift-get-nodes /etc/swift/container.ring.gz portal_0CE571BC7F1B4620B506 EDBF69825052593DDF2537E3 |less | 16:23 |
| eckesicle__ | and the output does not exist in the expected location on the server and disk specified | 16:24 |
| eckesicle__ | from the out | 16:24 |
| eckesicle__ | https://gist.github.com/cfeckardt/562d5281ea14d520a56d10b6bf67ce5c | 16:24 |
| *** mvk has quit IRC | 16:32 | |
| *** oshritf has quit IRC | 16:40 | |
| *** diogogmt has joined #openstack-swift | 16:41 | |
| *** garyj has joined #openstack-swift | 16:42 | |
| *** chlong has quit IRC | 16:45 | |
| *** jordanP has quit IRC | 16:46 | |
| *** rcernin has joined #openstack-swift | 16:46 | |
| *** tongli has quit IRC | 16:48 | |
| *** diogogmt has quit IRC | 16:49 | |
| acoles | eckesicle__: if the container doesn't exist then an object GET will default to using policy 0. If the original PUT was to a different policy then the object GET will 404. | 16:49 |
| *** diogogmt has joined #openstack-swift | 16:49 | |
| notmyname | good morning | 16:50 |
| acoles | eckesicle__: swiftclient will create a container when putting an object, so if your user uses that client or one with similar behaviour then that might explain how they managed to PUT and GET new objects. But that's just a guess on my part. | 16:51 |
| acoles | notmyname: good morning | 16:51 |
| acoles | eckesicle__: also note that any non-default policy must be specified on swift-get-nodes command line in order to get correct results | 16:53 |
| acoles | I know that winter solstice is approaching when it is dark before notmyname says good morning | 16:54 |
| notmyname | :-) | 16:54 |
| jrichli | notmyname: how was the conference last week? Any interesting outcomes? | 16:56 |
| *** chlong has joined #openstack-swift | 16:58 | |
| *** garyj has quit IRC | 16:59 | |
| *** diogogmt has quit IRC | 17:00 | |
| *** garyj has joined #openstack-swift | 17:00 | |
| notmyname | jrichli: it was good. someone found this and linked to it https://www.linkedin.com/pulse/value-attending-small-industry-events-jonathan-gershater?trk=hp-feed-article-title-comment | 17:01 |
| notmyname | the main benefit, as always, was talking to people. mordred and I had a good talk for a couple of hours. | 17:03 |
| notmyname | also, my wife was able to come with me, and that was nice | 17:03 |
| *** garyj has quit IRC | 17:04 | |
| *** mingyu has quit IRC | 17:05 | |
| notmyname | this is so cool. I just heard something really neat from a support person who sits next to me. good "swift is awesome" story | 17:06 |
| notmyname | so they've got a 4 node cluster. they had to do some power maintenance on 2 of them, so they'd be offline for ~15 hours or so. then, they have some automated reboot set up on a third one | 17:07 |
| notmyname | so for some period of time on an actively used cluster, they'll have 75% of the cluster down | 17:07 |
| notmyname | zero issues. everything worked just fine | 17:07 |
| *** vint_bra has quit IRC | 17:07 | |
| *** vint_bra has joined #openstack-swift | 17:07 | |
| notmyname | now, of course I'd expect swift to do well ;-) | 17:11 |
| notmyname | but there's not many storage systems that keep working when you've got 75% of the cluster down | 17:11 |
| *** vint_bra has quit IRC | 17:12 | |
| *** garyj has joined #openstack-swift | 17:13 | |
| notmyname | feature branch creation is now handled via gerrit reviews http://git.openstack.org/cgit/openstack/releases/tree/README.rst#n69 | 17:15 |
| jrichli | nice! (re: trip and swift survival story) | 17:16 |
| *** diogogmt has joined #openstack-swift | 17:17 | |
| jrichli | I think that crowd is a fire hazard | 17:18 |
| zaitcev | mathiasb: Use X-Auth-Token. Some implementations don't even look for X-Storage-Token anymore (e.g. Ceph RGW does that -- although amazingly enough its TempAuth actually _returns_ it for some reason). | 17:21 |
| *** tqtran has joined #openstack-swift | 17:21 | |
| mathiasb | zaitcev: thanks for the info! | 17:33 |
| *** jlwihte has joined #openstack-swift | 17:33 | |
| *** jlwihte has quit IRC | 17:33 | |
| *** jlwhite_ has joined #openstack-swift | 17:34 | |
| mathiasb | jrichli acoles: do you guys think you might have time to look at patch 364878 (again) before the holidays? | 17:37 |
| patchbot | https://review.openstack.org/#/c/364878/ - swift - Storing encryption root secret in Barbican | 17:37 |
| notmyname | Boston summit/forum openstack event call for papers is now open https://www.openstack.org/summit/boston-2017/call-for-presentations/ | 17:37 |
| *** JimCheung has joined #openstack-swift | 17:42 | |
| jrichli | mathiasb: Technically, my holidays start COB this Friday. :-) But I do *plan* to do some work over the break. I will put this higher on the list. I'll first have to setup keystone and barbican :-) might have some questions | 17:43 |
| mathiasb | jrichli: I'll be around until Friday next week to answer any and all questions you may have! :) | 17:45 |
| mathiasb | there was some good feedback already from the barbican folks, so hopefully the castellan/barbican usage issues are now mostly ironed out | 17:45 |
| acoles | mathiasb: I'll try. sorry, I know it's been there for a while. | 17:45 |
| *** abccba321 has joined #openstack-swift | 17:45 | |
| *** abccba321 has left #openstack-swift | 17:47 | |
| *** asettle has quit IRC | 17:48 | |
| garyj | jrichli: hello janie! | 17:48 |
| jrichli | hello, garyj | 17:48 |
| garyj | jrichli: this is Gary Jin from IBM. :) | 17:48 |
| *** rledisez has quit IRC | 17:49 | |
| *** diogogmt has quit IRC | 17:49 | |
| *** dmorita has joined #openstack-swift | 17:50 | |
| jrichli | Ah, yes. I didn't make the connection before that it was you in the recent convo about EC | 17:50 |
| jrichli | garyj ^^ | 17:50 |
| *** vint_bra has joined #openstack-swift | 17:52 | |
| *** klrmn has joined #openstack-swift | 17:54 | |
| _david_cole_soho | acoles: Thanks for your input. I will keep that in mind | 17:56 |
| garyj | :) yep, that's me. Now EC is bread and butter at the new gig. Did meet notmyname in SLC, very helpful. :) | 17:56 |
| jrichli | good to hear | 17:57 |
| *** Jeffrey4l has quit IRC | 17:58 | |
| *** asettle has joined #openstack-swift | 18:04 | |
| *** dmorita has quit IRC | 18:05 | |
| jrichli | garyj: what's the new gig, may I ask? | 18:06 |
| tdasilva | notmyname: i'm still trying to understand what's the format of the 'forum' ? is it more like open discussion (design summit) or presentations? | 18:07 |
| notmyname | tdasilva: from what I can tell, it's the conference part of the old summit. so the marketing expo hall, the "stand at the front of the room" talks. and the ops sessions | 18:07 |
| notmyname | the design session part is the PTG | 18:08 |
| garyj | jrichli: i am with a startup called phazr.io building a new ec tech that integrates into swift - an alternative to jerasure | 18:08 |
| jrichli | oh, right! very exciting | 18:09 |
| timburke | good morning | 18:13 |
| garyj | jrichli: you should see some code from us coming in the next few weeks, probably sometime in Jan. | 18:13 |
| jrichli | great! | 18:14 |
| * jrichli is gonna grab some lunch | 18:14 | |
| *** dmorita has joined #openstack-swift | 18:15 | |
| *** asettle has quit IRC | 18:17 | |
| *** acoles is now known as acoles_ | 18:21 | |
| *** david-lyle has quit IRC | 18:21 | |
| *** vinsh has joined #openstack-swift | 18:29 | |
| *** david-lyle has joined #openstack-swift | 18:30 | |
| notmyname | ugh. that realization that "the next few weeks" means january ;-) | 18:34 |
| *** ukaynar_ has joined #openstack-swift | 18:34 | |
| garyj | notmyname: tick-tock-tick-tock :) | 18:34 |
| *** ukaynar_ has quit IRC | 18:39 | |
| notmyname | any comments on https://review.openstack.org/#/c/409924/ today would be helpful. I'd like to get it released this week | 18:40 |
| patchbot | patch 409924 - swift - Swift 2.12.0 authors/changelog updates | 18:40 |
| notmyname | acoles_: jrichli: mattoliverau: tdasilva: zaitcev: clayg: ^ | 18:40 |
| *** ukaynar has joined #openstack-swift | 18:48 | |
| *** tsg has quit IRC | 18:50 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: Support Read/Write Affinity Per Policy https://review.openstack.org/382766 | 19:01 |
| *** ukaynar has quit IRC | 19:06 | |
| *** ukaynar has joined #openstack-swift | 19:09 | |
| zaitcev | "Remove empty db hash and suffix directories if a db gets quarantined." - sounds like a dangerous and useless thing to do.... If DB was there, there's going to be one again very soon, so why bother removing | 19:10 |
| zaitcev | I don't remember that review | 19:10 |
| *** amoralej is now known as amoralej|off | 19:11 | |
| *** ukaynar has quit IRC | 19:12 | |
| zaitcev | Bug 1583719 is reported by Clay, hmm.... | 19:12 |
| openstack | bug 1583719 in OpenStack Object Storage (swift) "empty db hash/suffix directories can pile up" [Medium,Fix released] https://launchpad.net/bugs/1583719 - Assigned to Mahati Chamarthy (mahati-chamarthy) | 19:12 |
| *** dmorita has quit IRC | 19:17 | |
| *** dmorita has joined #openstack-swift | 19:18 | |
| *** joeljwright has quit IRC | 19:19 | |
| *** siva_krish has quit IRC | 19:22 | |
| *** dmorita has quit IRC | 19:24 | |
| *** McMurlock1 has joined #openstack-swift | 19:24 | |
| *** dmorita has joined #openstack-swift | 19:25 | |
| *** dmorita has quit IRC | 19:25 | |
| *** siva_krish has joined #openstack-swift | 19:25 | |
| *** McMurlock1 has quit IRC | 19:26 | |
| *** McMurlock1 has joined #openstack-swift | 19:27 | |
| *** cbartz has joined #openstack-swift | 19:27 | |
| *** dmorita has joined #openstack-swift | 19:27 | |
| *** McMurlock1 has quit IRC | 19:27 | |
| *** geaaru has quit IRC | 19:29 | |
| *** joeljwright has joined #openstack-swift | 19:29 | |
| *** ChanServ sets mode: +v joeljwright | 19:29 | |
| jrichli | zaitcev: you might be interested in some convo that happened about that patch - see eavesdrop on 2016-12-01T21:47:09, 2016-12-01T22:51:51, 2016-12-02T04:16:30 | 19:31 |
| *** joeljwright has quit IRC | 19:34 | |
| notmyname | tdasilva: zaitcev: thanks for looking at the changelog. pushing the yaml version now | 19:39 |
| openstackgerrit | John Dickinson proposed openstack/swift: Swift 2.12.0 authors/changelog updates https://review.openstack.org/409924 | 19:40 |
| notmyname | unless something else comes up soon, i'll land it after lunch | 19:40 |
| *** joeljwright has joined #openstack-swift | 19:49 | |
| *** ChanServ sets mode: +v joeljwright | 19:49 | |
| *** oshritf has joined #openstack-swift | 19:55 | |
| *** ChubYann has joined #openstack-swift | 20:05 | |
| *** siva_krish has quit IRC | 20:07 | |
| *** oshritf has quit IRC | 20:07 | |
| *** siva_krish has joined #openstack-swift | 20:10 | |
| *** dmorita has quit IRC | 20:10 | |
| *** joeljwright has quit IRC | 20:11 | |
| *** dmorita has joined #openstack-swift | 20:14 | |
| *** dfflanders has joined #openstack-swift | 20:15 | |
| *** asettle has joined #openstack-swift | 20:18 | |
| *** dmorita has quit IRC | 20:18 | |
| *** asettle has quit IRC | 20:22 | |
| *** garyj has quit IRC | 20:32 | |
| *** pcaruana has quit IRC | 20:33 | |
| *** StevenK has quit IRC | 20:38 | |
| *** StevenK has joined #openstack-swift | 20:38 | |
| *** clu_ has joined #openstack-swift | 20:40 | |
| *** diogogmt has joined #openstack-swift | 20:46 | |
| *** mvk has joined #openstack-swift | 20:46 | |
| *** joeljwright has joined #openstack-swift | 20:47 | |
| *** ChanServ sets mode: +v joeljwright | 20:47 | |
| openstackgerrit | John Dickinson proposed openstack/swift: Swift 2.12.0 authors/changelog updates https://review.openstack.org/409924 | 20:50 |
| notmyname | timburke: fixed the s/users/user/ typo | 20:51 |
| notmyname | ...and landing it | 20:51 |
| *** acoles_ is now known as acoles | 20:53 | |
| notmyname | meeting time in #openstack-meeting | 20:57 |
| notmyname | well, in 2.5 minute s:-) | 20:57 |
| acoles | got me all excited and now I have to wait | 20:58 |
| mattoliverau | morning | 20:59 |
| notmyname | I know this is the highlight of your week acoles | 20:59 |
| joeljwright | :) | 21:00 |
| notmyname | ....now! | 21:00 |
| jrichli | mattoliverau: thanks for introducing me to tonyb last night, btw. hi tonyb! | 21:00 |
| acoles | I don't get out much :) | 21:00 |
| jrichli | lol | 21:00 |
| mattoliverau | jrichli: your welcome :) | 21:01 |
| *** dmorita has joined #openstack-swift | 21:01 | |
| *** cbartz has quit IRC | 21:08 | |
| *** m_kazuhiro has joined #openstack-swift | 21:10 | |
| *** ukaynar has joined #openstack-swift | 21:11 | |
| *** ukaynar_ has joined #openstack-swift | 21:17 | |
| *** nikivi has joined #openstack-swift | 21:18 | |
| *** ukaynar has quit IRC | 21:19 | |
| *** asettle has joined #openstack-swift | 21:21 | |
| *** nikivi has quit IRC | 21:22 | |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: Fix up some formatting https://review.openstack.org/410959 | 21:23 |
| *** catintheroof has joined #openstack-swift | 21:24 | |
| *** asettle has quit IRC | 21:25 | |
| openstackgerrit | Joel Wright proposed openstack/swift: Add Preamble and Postamble to SLO and SegmentedIterable https://review.openstack.org/365371 | 21:26 |
| *** garyj has joined #openstack-swift | 21:32 | |
| openstackgerrit | Joel Wright proposed openstack/swift: Add Preamble and Postamble to SLO and SegmentedIterable https://review.openstack.org/365371 | 21:33 |
| tdasilva | clayg: hi, did you get any inputs from pypi guys about that issue with the bugtracker metadata? | 21:33 |
| clayg | nope | 21:34 |
| clayg | well that's not true | 21:34 |
| clayg | one guy said it's not possible to fix it | 21:34 |
| tdasilva | lol | 21:35 |
| tdasilva | that sounds horrible | 21:36 |
| tdasilva | and hard to believe | 21:36 |
| *** garyj has quit IRC | 21:37 | |
| *** lespaul has joined #openstack-swift | 21:40 | |
| *** dja has joined #openstack-swift | 21:40 | |
| *** lespaul has quit IRC | 21:46 | |
| *** jlwhite_ has quit IRC | 21:47 | |
| *** garyj has joined #openstack-swift | 21:55 | |
| *** garyj has quit IRC | 21:59 | |
| *** sams-gleb has quit IRC | 22:00 | |
| clayg | tdasilva: I think pypi is in the process of a rewrite | 22:00 |
| tdasilva | clayg: currently on centos we have 4 or 5 swift rpms | 22:00 |
| tdasilva | from one repo | 22:00 |
| notmyname | clayg: tdasilva: what's the pypi bugtracker metadata issue? | 22:00 |
| clayg | tdasilva: they've done some trimming of features in order to triage things until the new stuff is depoyed | 22:01 |
| *** sams-gleb has joined #openstack-swift | 22:01 | |
| kota_ | so if we have several ways to make language binding, it may solve the packaging problem? | 22:01 |
| tdasilva | notmyname: https://pypi.python.org/pypi/PyECLib/1.4.0 | 22:01 |
| kota_ | likely setup.py for python and Makefile for c and something for golang? | 22:01 |
| clayg | tdasilva: I checked out the source and there's some code that used to update it - but I sort of believe it may infact be current inaccessable :'( | 22:01 |
| notmyname | tdasilva: ah, right | 22:01 |
| clayg | tdasilva: maybe if we ask dstufft really nicely he can just find some with access to update the data store directly | 22:01 |
| kota_ | and each setupper works only for making each lang's binary file. | 22:01 |
| clayg | kota_: probably the makefile would call the stup.py but yeah something like that | 22:02 |
| *** m_kazuhiro has quit IRC | 22:02 | |
| clayg | kota_: mostly just that liberasurecode *could* crap out a dynamic shared library but by default pyeclib would mostly just build a thing that could be used directly | 22:02 |
| tdasilva | also...does tsg and kevin care if we joing the repos? | 22:03 |
| tdasilva | i'm guessing not | 22:03 |
| clayg | tdasilva: idk, they thought originally it made sense for them to be seperate | 22:03 |
| clayg | I think keeping the CPython code a think wrapper around a good solid libec interface is still important | 22:03 |
| clayg | having the CPython code + setup.py in the libec source tree probably doesn't matter | 22:04 |
| kota_ | clayg: no no no, if we calls the setupper (i.e. make file) for c lang, just building shared library for that | 22:04 |
| clayg | someone might think that the nature of pyeclib using whatever dynamic liberassurecode .so it finds at run time had some practical utility - but for me it mostly just sucks | 22:04 |
| *** ukaynar_ has quit IRC | 22:04 | |
| kota_ | clayg: make file also can build non-shared lib (i.e. .a) so that, for python setupper (i.e. setup.py) will call make file to make .a file and includes the lib into the c binary for python? | 22:05 |
| *** sams-gleb has quit IRC | 22:05 | |
| clayg | kota_: no sorry - i just meant that makefile's are pretty flexible - it can have many tasks - one would be to build the python code for us | 22:05 |
| clayg | kota_: I don't really know how that works but I like everyhting you said | 22:06 |
| onovy | please use setup.py for "building" python code | 22:06 |
| clayg | *sigh* | 22:06 |
| onovy | make python can call "python setup.py ..." | 22:06 |
| clayg | nm, forget i said anything | 22:06 |
| *** joeljwright has quit IRC | 22:06 | |
| tdasilva | lol | 22:06 |
| tdasilva | sounds like you all saying the same thing | 22:06 |
| tdasilva | i need to step out now, will be back in about 3hrs to talk to m_kazuhiro about auto-tiering, if anyone wants to participate let me know | 22:07 |
| kota_ | tdasilva: great! | 22:07 |
| clayg | tdasilva: thanks | 22:07 |
| * kota_ is also step out to find breakfast and go office early | 22:08 | |
| *** acoles is now known as acoles_ | 22:13 | |
| *** catintheroof has quit IRC | 22:19 | |
| *** arch-nemesis has joined #openstack-swift | 22:21 | |
| *** nikivi has joined #openstack-swift | 22:31 | |
| *** nikivi has quit IRC | 22:36 | |
| *** nikivi has joined #openstack-swift | 22:42 | |
| *** nikivi has quit IRC | 22:45 | |
| eckesicle__ | im seeing another interesting error in my logs | 22:54 |
| eckesicle__ | https://www.irccloud.com/pastebin/DiZwFVDf/ | 22:54 |
| eckesicle__ | "Deferring reject downstream" | 22:54 |
| eckesicle__ | Is that something to be concerned about? | 22:55 |
| eckesicle__ | I tried to grep the sauce for that string, but nothing came up | 22:55 |
| *** garyj has joined #openstack-swift | 22:55 | |
| clayg | sound's keystone-y to me | 22:55 |
| timburke | that's coming out of keystonemiddleware's auth_token middleware. nothing to worry about | 22:57 |
| eckesicle__ | ah fab | 22:58 |
| eckesicle__ | there's also a lot of inactivity in the syslog | 22:58 |
| eckesicle__ | on every server | 22:58 |
| eckesicle__ | HEADs and PUTs | 22:58 |
| eckesicle__ | (we recently rebuilt the ring files) | 22:58 |
| eckesicle__ | that just means it's moving files to their new locations? | 22:58 |
| *** dmorita has quit IRC | 23:03 | |
| *** rcernin has quit IRC | 23:05 | |
| * kota_ is back to online at office | 23:08 | |
| clayg | eckesicle__: EC or repliated data? | 23:08 |
| eckesicle__ | what does that mean? | 23:09 |
| tonyb | jrichli: o/ | 23:10 |
| *** dmorita has joined #openstack-swift | 23:10 | |
| clayg | on a ring rebalance, a storage policy that uses erasure coded data would introduce a bunch of GETs and some PUTs (but I think only the GETs a re logged really) - I wouldn't expect replicated data to change the number of HEAD or PUT requests you see logged | 23:11 |
| clayg | ... if that's what you're looking at? | 23:11 |
| *** dmorita has quit IRC | 23:12 | |
| *** dmorita has joined #openstack-swift | 23:12 | |
| eckesicle__ | hmm | 23:13 |
| eckesicle__ | okay i will investigate it a bit further | 23:13 |
| eckesicle__ | (the underlying issue is that some files appear inaccessible for some users) | 23:13 |
| eckesicle__ | and i suspect they're not on the correct hosts | 23:16 |
| *** asettle has joined #openstack-swift | 23:22 | |
| clayg | eckesicle__: a single rebalance by itself would never cause an availability issue - must be more going on - good luck | 23:24 |
| eckesicle__ | cheers! | 23:24 |
| eckesicle__ | it's very strange | 23:24 |
| eckesicle__ | i suspect the files simply don't exist | 23:24 |
| eckesicle__ | but are still listed in the containers | 23:24 |
| eckesicle__ | that's my working theory anyways | 23:25 |
| *** asettle has quit IRC | 23:27 | |
| *** chlong has quit IRC | 23:37 | |
| *** klamath has quit IRC | 23:46 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
