| mattoliverau | notmyname: The LCA call for papers closes reletively soon, You submitted yours? (I work with the chair's of the paper commitee, and he keeps reminding the team, so why not pass on the love) :P Besides what's an LCA without a talk from notmyname :) | 00:16 |
|---|---|---|
| notmyname | heh | 00:23 |
| notmyname | not yet, but it's on my radar | 00:23 |
| notmyname | I gotta talk to the boss man about it. LCA is not quite the same as a conference in denter/portland/atlanta | 00:24 |
| notmyname | also *denver | 00:25 |
| mattoliverau | lol | 00:25 |
| *** dmorita has joined #openstack-swift | 00:26 | |
| mattoliverau | notmyname: yeah understandable, let me know if you can make it, if not, I'll submit a proposal to talk SP etc, so we can keep swift as a continuing trending topic ;) .. but coming from me wont be the same as coming from the PTL :) | 00:28 |
| notmyname | mattoliverau: I believe that peluse_ is going to be submitting a patch | 00:28 |
| mattoliverau | peluse_: awesome, I'll leave it in his capible hands then :) | 00:29 |
| mattoliverau | notmyname i mean ^^ | 00:29 |
| *** patchbot has quit IRC | 00:36 | |
| *** patchbot has joined #openstack-swift | 00:38 | |
| *** patchbot has quit IRC | 00:38 | |
| peluse_ | mattoliverau: yup, I'll be submitting a proposal or 2 for LCA. Policies is of course really big, Erasure Coding is coming and that can be sliced and diced a bunch of different ways... something for sure though! | 00:41 |
| mattoliverau | peluse_: awesome! It'll be good to have you down this way! | 00:46 |
| *** shri has quit IRC | 00:56 | |
| *** patchbot has joined #openstack-swift | 01:01 | |
| notmyname | torgomatic: suppose we wanted to talk about patch 103731. It would be nice if we had an automatic link to it | 01:02 |
| patchbot | notmyname: https://review.openstack.org/#/c/103731/ | 01:02 |
| clayg | and what about patch 999999 | 01:06 |
| patchbot | clayg: https://review.openstack.org/#/c/999999/ | 01:06 |
| clayg | oic | 01:06 |
| clayg | even 888888 | 01:07 |
| clayg | quite nice :) | 01:07 |
| clayg | patches are great 777777 | 01:07 |
| clayg | hrmmm... | 01:07 |
| patchbot | clayg: You've given me 5 invalid commands within the last minute; I'm now ignoring you for 10 minutes. | 01:07 |
| notmyname | clayg: but if you talke about patch xyzzy, nothing happens | 01:07 |
| notmyname | lol | 01:07 |
| clayg | now that is a FEATURE! | 01:07 |
| notmyname | hmm,, but why was it triggered? | 01:08 |
| notmyname | another line? | 01:08 |
| notmyname | ignore me too! | 01:08 |
| clayg | i guess he thinks I'm still talking to him? | 01:08 |
| clayg | patchbot: are you *really* ignoring me?! | 01:08 |
| clayg | awwww :'( | 01:08 |
| notmyname | INFO 2014-07-02T01:12:27 Ignoring | 01:09 |
| notmyname | *!~clayg@ec2-54-235-44-253.compute-1.amazonaws.com for 600 seconds due to | 01:09 |
| notmyname | an apparent invalid command flood. | 01:09 |
| notmyname | but why did it think you did commands? | 01:09 |
| notmyname | is it the numbers, lke 23423424 | 01:09 |
| notmyname | and 2345 | 01:10 |
| notmyname | and 789 | 01:10 |
| notmyname | and 1233 | 01:10 |
| notmyname | nopw | 01:10 |
| notmyname | nope | 01:10 |
| notmyname | patchbot doesn't like clayg | 01:10 |
| *** mitz_ has joined #openstack-swift | 01:11 | |
| clayg | i got skillz | 01:11 |
| notmyname | haxor skillz | 01:11 |
| zaitcev | nice | 01:11 |
| notmyname | today's IRC flood comes to you courtesy of playing with new toys | 01:12 |
| clayg | lolwut "In the endeavor to move from the default deployment of Keystone being eventlet (in devstack) to Apache + mod_wsgi" | 01:19 |
| zaitcev | I don't remember where it comes from but Certain People were plugging at it for a while, with a preservance and dilligence that deserved a better application. | 01:21 |
| clayg | no, i may be wrong on both counts, mod_python seems to have gotten a new maintainer as well | 01:21 |
| *** kota_ has joined #openstack-swift | 01:36 | |
| *** ho has joined #openstack-swift | 01:43 | |
| *** zaitcev has quit IRC | 01:46 | |
| *** nosnos has joined #openstack-swift | 01:49 | |
| *** nosnos has quit IRC | 01:55 | |
| *** nosnos has joined #openstack-swift | 01:55 | |
| *** nsquare has quit IRC | 01:59 | |
| *** nsquare has joined #openstack-swift | 02:00 | |
| *** zaitcev has joined #openstack-swift | 02:03 | |
| *** ChanServ sets mode: +v zaitcev | 02:03 | |
| *** acoles has quit IRC | 02:13 | |
| *** nsquare has quit IRC | 02:17 | |
| *** zhiyan_ is now known as zhiyan | 02:22 | |
| *** acoles has joined #openstack-swift | 02:22 | |
| *** ChanServ sets mode: +v acoles | 02:22 | |
| *** zaitcev has quit IRC | 02:25 | |
| *** zhiyan is now known as zhiyan_ | 02:30 | |
| *** zhiyan_ is now known as zhiyan | 02:42 | |
| *** nsquare has joined #openstack-swift | 02:48 | |
| *** ajc_ has joined #openstack-swift | 02:50 | |
| *** theannegentle has joined #openstack-swift | 03:03 | |
| *** theannegentle has quit IRC | 03:04 | |
| *** gyee has quit IRC | 03:10 | |
| *** zhiyan is now known as zhiyan_ | 03:22 | |
| *** zhiyan_ is now known as zhiyan | 03:24 | |
| *** nosnos has quit IRC | 03:27 | |
| *** zhiyan has quit IRC | 03:30 | |
| *** zhiyan has joined #openstack-swift | 03:37 | |
| *** zhiyan is now known as zhiyan_ | 03:39 | |
| *** zhiyan_ is now known as zhiyan | 03:40 | |
| *** zhiyan is now known as zhiyan_ | 03:46 | |
| *** zhiyan_ is now known as zhiyan | 03:50 | |
| *** ky has joined #openstack-swift | 04:01 | |
| *** km has joined #openstack-swift | 04:01 | |
| *** km is now known as Guest40940 | 04:02 | |
| *** tong_ has joined #openstack-swift | 04:02 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:03 | |
| *** tongli has quit IRC | 04:03 | |
| *** kota_ has quit IRC | 04:05 | |
| *** kota_ has joined #openstack-swift | 04:05 | |
| *** ktsuyuzaki has quit IRC | 04:07 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:08 | |
| *** kota_ has quit IRC | 04:10 | |
| *** kota_ has joined #openstack-swift | 04:10 | |
| *** ktsuyuzaki has quit IRC | 04:12 | |
| *** ho has quit IRC | 04:13 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:13 | |
| anticw | 12345 | 04:14 |
| *** kota_ has quit IRC | 04:15 | |
| *** kota_ has joined #openstack-swift | 04:17 | |
| *** ktsuyuzaki has quit IRC | 04:20 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:20 | |
| *** kota_ has quit IRC | 04:22 | |
| *** kota_ has joined #openstack-swift | 04:23 | |
| *** ktsuyuzaki has quit IRC | 04:24 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:25 | |
| *** kota_ has quit IRC | 04:27 | |
| *** ktsuyuza_ has joined #openstack-swift | 04:28 | |
| *** ktsuyuzaki has quit IRC | 04:30 | |
| *** kota_ has joined #openstack-swift | 04:30 | |
| *** ktsuyuza_ has quit IRC | 04:32 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:33 | |
| *** tong_ has quit IRC | 04:33 | |
| *** kota_ has quit IRC | 04:35 | |
| *** kota_ has joined #openstack-swift | 04:37 | |
| *** ktsuyuzaki has quit IRC | 04:39 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:40 | |
| *** nosnos has joined #openstack-swift | 04:42 | |
| *** kota_ has quit IRC | 04:42 | |
| *** kota_ has joined #openstack-swift | 04:42 | |
| *** ktsuyuzaki has quit IRC | 04:44 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:45 | |
| *** kota_ has quit IRC | 04:47 | |
| *** kota_ has joined #openstack-swift | 04:47 | |
| *** ktsuyuzaki has quit IRC | 04:49 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:50 | |
| *** kota_ has quit IRC | 04:52 | |
| *** ppai has joined #openstack-swift | 04:52 | |
| *** kota_ has joined #openstack-swift | 04:52 | |
| *** psharma has joined #openstack-swift | 04:53 | |
| *** ktsuyuzaki has quit IRC | 04:54 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:57 | |
| *** kota_ has quit IRC | 04:59 | |
| *** kota_ has joined #openstack-swift | 05:01 | |
| *** ktsuyuzaki has quit IRC | 05:03 | |
| *** ktsuyuzaki has joined #openstack-swift | 05:06 | |
| *** kota_ has quit IRC | 05:08 | |
| *** kota_ has joined #openstack-swift | 05:08 | |
| *** Midnightmyth has joined #openstack-swift | 05:10 | |
| *** ktsuyuzaki has quit IRC | 05:10 | |
| *** ktsuyuzaki has joined #openstack-swift | 05:13 | |
| *** mmcardle has joined #openstack-swift | 05:14 | |
| *** kota_ has quit IRC | 05:15 | |
| *** kota_ has joined #openstack-swift | 05:16 | |
| *** ktsuyuzaki has quit IRC | 05:18 | |
| *** ktsuyuzaki has joined #openstack-swift | 05:18 | |
| *** kota_ has quit IRC | 05:20 | |
| *** kota_ has joined #openstack-swift | 05:21 | |
| *** ktsuyuzaki has quit IRC | 05:23 | |
| *** ktsuyuzaki has joined #openstack-swift | 05:23 | |
| *** kota_ has quit IRC | 05:25 | |
| *** kota_ has joined #openstack-swift | 05:26 | |
| *** swat30 has quit IRC | 05:26 | |
| *** zul has quit IRC | 05:27 | |
| *** swat30 has joined #openstack-swift | 05:27 | |
| *** ktsuyuzaki has quit IRC | 05:27 | |
| *** zul has joined #openstack-swift | 05:28 | |
| *** ktsuyuzaki has joined #openstack-swift | 05:30 | |
| *** kota_ has quit IRC | 05:32 | |
| *** kota_ has joined #openstack-swift | 05:33 | |
| *** ktsuyuzaki has quit IRC | 05:35 | |
| *** ktsuyuza_ has joined #openstack-swift | 05:35 | |
| *** chandan_kumar has joined #openstack-swift | 05:38 | |
| *** kota_ has quit IRC | 05:38 | |
| *** ktsuyuzaki has joined #openstack-swift | 05:38 | |
| *** ktsuyuza_ has quit IRC | 05:40 | |
| *** bkopilov_ has quit IRC | 05:44 | |
| *** bkopilov has joined #openstack-swift | 05:44 | |
| *** kota_ has joined #openstack-swift | 05:44 | |
| *** ktsuyuzaki has quit IRC | 05:47 | |
| *** ktsuyuzaki has joined #openstack-swift | 05:47 | |
| *** kota_ has quit IRC | 05:49 | |
| *** kota_ has joined #openstack-swift | 05:51 | |
| *** ktsuyuzaki has quit IRC | 05:53 | |
| *** ktsuyuzaki has joined #openstack-swift | 05:54 | |
| *** kota_ has quit IRC | 05:56 | |
| *** ktsuyuzaki has quit IRC | 06:17 | |
| *** kota_ has joined #openstack-swift | 06:17 | |
| *** Midnightmyth has quit IRC | 06:30 | |
| *** ky has quit IRC | 06:39 | |
| *** mitz_ has quit IRC | 06:48 | |
| *** mitz_ has joined #openstack-swift | 06:50 | |
| *** nsquare has quit IRC | 06:51 | |
| *** yuan has quit IRC | 07:00 | |
| *** yuan has joined #openstack-swift | 07:00 | |
| *** yuan has quit IRC | 07:00 | |
| *** yuan has joined #openstack-swift | 07:00 | |
| *** foexle has joined #openstack-swift | 07:04 | |
| *** jamiehannaford has joined #openstack-swift | 07:04 | |
| *** foexle has quit IRC | 07:20 | |
| openstackgerrit | Matthew Oliver proposed a change to openstack/swift: Add X-Delete-At/After support to FormPost https://review.openstack.org/104101 | 07:28 |
| openstackgerrit | Matthew Oliver proposed a change to openstack/swift: Add X-Delete-At/After support to FormPost https://review.openstack.org/104101 | 07:32 |
| mattoliverau | On that note, time to go make dinner :) | 07:34 |
| *** mkerrin has joined #openstack-swift | 07:45 | |
| *** mmcardle has joined #openstack-swift | 07:52 | |
| *** mmcardle has quit IRC | 08:02 | |
| *** nacim has joined #openstack-swift | 08:02 | |
| *** mmcardle has joined #openstack-swift | 08:07 | |
| *** foexle has joined #openstack-swift | 08:07 | |
| *** mkollaro has joined #openstack-swift | 08:14 | |
| *** fifieldt_ has joined #openstack-swift | 08:16 | |
| *** fifieldt has quit IRC | 08:18 | |
| *** andyandy has joined #openstack-swift | 08:21 | |
| *** dANOKELOFF has joined #openstack-swift | 08:22 | |
| dANOKELOFF | Hi , in swift with curl how we can destroy a non-empty containers ? | 08:25 |
| omame | you can't | 08:25 |
| omame | you need to delete all the objects first, then you can delete the container | 08:26 |
| dANOKELOFF | omame: oh... ok , and curl command exist for delete all objects? | 08:26 |
| dANOKELOFF | omame: or i need to delete object one-by-one | 08:26 |
| omame | the latter | 08:27 |
| omame | or use swiftly | 08:27 |
| omame | check out the --recursive flag at http://greg.brim.net/swiftly/2.04/#swiftly-help-delete | 08:27 |
| dANOKELOFF | omame: ok , thank's | 08:29 |
| *** fifieldt_ is now known as fifieldt | 08:31 | |
| *** mkollaro has quit IRC | 08:35 | |
| *** jamiehannaford has quit IRC | 08:44 | |
| *** jamiehannaford has joined #openstack-swift | 08:44 | |
| *** bvandenh has joined #openstack-swift | 08:48 | |
| *** jamiehannaford has quit IRC | 08:52 | |
| *** Guest40940 has quit IRC | 08:55 | |
| *** jamiehannaford has joined #openstack-swift | 08:56 | |
| *** dmorita has quit IRC | 09:11 | |
| *** mkollaro has joined #openstack-swift | 09:19 | |
| *** gvernik has joined #openstack-swift | 09:22 | |
| *** kota_ has quit IRC | 09:23 | |
| *** ujjain has quit IRC | 09:42 | |
| *** gvernik has quit IRC | 09:44 | |
| *** nacim has quit IRC | 09:48 | |
| *** dANOKELOFF has quit IRC | 09:55 | |
| *** nacim has joined #openstack-swift | 09:58 | |
| *** ppai has quit IRC | 10:04 | |
| *** ppai has joined #openstack-swift | 10:20 | |
| *** mmcardle has quit IRC | 10:22 | |
| *** ujjain has joined #openstack-swift | 10:23 | |
| *** ujjain has quit IRC | 10:24 | |
| *** ujjain has joined #openstack-swift | 10:24 | |
| *** mmcardle has joined #openstack-swift | 10:26 | |
| *** ppai has quit IRC | 10:46 | |
| *** TaiSHi has joined #openstack-swift | 10:50 | |
| *** mkollaro has quit IRC | 11:05 | |
| *** lsell has left #openstack-swift | 11:18 | |
| *** Midnightmyth has joined #openstack-swift | 11:33 | |
| *** ppai has joined #openstack-swift | 11:39 | |
| *** mkollaro has joined #openstack-swift | 11:46 | |
| *** mmcardle has quit IRC | 11:55 | |
| *** ajc_ has quit IRC | 11:58 | |
| *** nosnos has quit IRC | 11:59 | |
| *** nosnos has joined #openstack-swift | 11:59 | |
| *** ppai has quit IRC | 12:02 | |
| *** nosnos has quit IRC | 12:04 | |
| *** Midnightmyth has quit IRC | 12:05 | |
| *** bvandenh has quit IRC | 12:08 | |
| *** sombrafam has joined #openstack-swift | 12:14 | |
| *** dANOKELOFF has joined #openstack-swift | 12:14 | |
| *** annegent_ has joined #openstack-swift | 12:15 | |
| dANOKELOFF | It's possible to create tempurl when your objects name is id/title . And your url is so : /account/containers/id/title ? | 12:15 |
| dANOKELOFF | Because id/title represent object | 12:15 |
| ahale | should work - if it doesn't it would be a bug | 12:16 |
| *** ppai has joined #openstack-swift | 12:18 | |
| *** Midnightmyth has joined #openstack-swift | 12:21 | |
| *** annegent_ is now known as annegentle_ | 12:21 | |
| *** nshaikh has joined #openstack-swift | 12:22 | |
| dANOKELOFF | I need to put my key in the tempurl filter ? | 12:25 |
| ahale | haven't set it up for a while, but from what i remember you set the key on the account in an X- header, and then include that when you build the hmac you provide to client to use | 12:29 |
| *** mmcardle has joined #openstack-swift | 12:30 | |
| dANOKELOFF | ok thank's | 12:31 |
| *** mlipchuk has joined #openstack-swift | 12:37 | |
| *** mlipchuk has quit IRC | 12:40 | |
| *** mlipchuk has joined #openstack-swift | 12:41 | |
| *** nshaikh has left #openstack-swift | 12:43 | |
| *** mmcardle has quit IRC | 12:47 | |
| *** mmcardle has joined #openstack-swift | 12:51 | |
| *** annegentle_ has quit IRC | 12:55 | |
| *** swat30 has quit IRC | 12:58 | |
| *** kenhui has joined #openstack-swift | 12:58 | |
| *** swat30 has joined #openstack-swift | 12:58 | |
| *** chandan_kumar is now known as chandankumar | 13:03 | |
| openstackgerrit | gholt proposed a change to openstack/swift: SimpleClient http proxying https://review.openstack.org/102598 | 13:08 |
| *** psharma has quit IRC | 13:09 | |
| dANOKELOFF | When i try to setup the temp_url : swift-temp-url GET 1200 /v1/AUTH_replication/7/48/f26af6ca-bf2c-48bd-bf45-3eac9a711bdc.mp3 testkey , that return tempurl but when i try to curl -i the url i have Error 401 Unauthorized : Temp URL invalid, anyone know this bug ? | 13:12 |
| *** ppai has quit IRC | 13:22 | |
| *** annegent_ has joined #openstack-swift | 13:31 | |
| *** Midnightmyth has quit IRC | 13:41 | |
| *** robsparker has joined #openstack-swift | 13:43 | |
| *** annegent_ has quit IRC | 13:43 | |
| *** annegent_ has joined #openstack-swift | 13:44 | |
| *** kenhui has quit IRC | 13:49 | |
| *** kenhui has joined #openstack-swift | 13:50 | |
| *** kenhui has quit IRC | 13:50 | |
| *** kenhui has joined #openstack-swift | 13:51 | |
| ctennis | dANOKELOFF: when you curl, did you include the entire tempurl in quotes? | 13:53 |
| dANOKELOFF | ctennis: i try but same error again | 13:55 |
| *** miqui has quit IRC | 13:56 | |
| *** bboris has joined #openstack-swift | 13:56 | |
| bboris | hello | 14:01 |
| bboris | can anyone tell me how many puts/second can i get with swift using swift all-in-one setup with 5 disks/zones and 3 replicas? currently i'm testing 50 parallel operations and i get results like 53 puts/s for 64k objects | 14:06 |
| bboris | is this normal? | 14:07 |
| *** mmcardle has quit IRC | 14:09 | |
| bboris | if anyone is interested in my results of ceph/swift tests, you can find them here: http://boris.descom.com/ceph-swift-140630.ods | 14:10 |
| *** mmcardle has joined #openstack-swift | 14:15 | |
| *** dANOKELOFF has quit IRC | 14:15 | |
| bboris | swift is quite faster in almost everything it seems, but compared to fopen writes per second... i am about to test writing to a single disk through nfs and compare that to swift, but i am waiting swift to finish replicating the objects of my last test, which was putting 4-32k objects for an hour, then get and delete. it's replicating for 18 hours now, and still going | 14:15 |
| *** kenhui has quit IRC | 14:25 | |
| *** mmcardle has quit IRC | 14:29 | |
| *** dANOKELOFF has joined #openstack-swift | 14:30 | |
| *** kenhui has joined #openstack-swift | 14:31 | |
| *** mmcardle has joined #openstack-swift | 14:32 | |
| *** zaitcev has joined #openstack-swift | 14:35 | |
| *** ChanServ sets mode: +v zaitcev | 14:35 | |
| *** thurloat has quit IRC | 14:36 | |
| *** mkollaro has quit IRC | 14:40 | |
| creiht | bboris: in your test, are you writing to a single container or multiple containers? | 14:42 |
| *** thurloat has joined #openstack-swift | 14:42 | |
| creiht | if you aren't using multiple containers that could be part of it | 14:43 |
| bboris | each thread creates it's own container in the beginning, writes as much objects as it can for 60 minutes, gets them back, deletes them and deletes the container | 14:43 |
| creiht | how many threads? | 14:43 |
| bboris | 50 | 14:43 |
| creiht | ahh | 14:43 |
| creiht | yeah you should be good there then | 14:44 |
| bboris | on a different machine, also | 14:44 |
| bboris | the network is good, i can reach almost 1 Gbit | 14:45 |
| creiht | have you tweaked any of your worker settings? | 14:45 |
| bboris | nothing, as is from the swift all-in-one documentation | 14:46 |
| creiht | k | 14:46 |
| creiht | everything defaults to only 1 worker with saio | 14:47 |
| creiht | when running your benchmark look at the swift processes, and see if any of them are reaching 100% cpu | 14:48 |
| creiht | like maybe the proxy for example | 14:48 |
| creiht | if they are then bump the worker count up | 14:48 |
| creiht | the saio is probably not the best thing to benchmark though | 14:49 |
| creiht | as it simulates running 4 servers on 1 | 14:49 |
| creiht | still interesting to test though :) | 14:49 |
| creiht | gotta run to the dentist | 14:50 |
| bboris | it's true, but i don't have free machines. also, for comparison, ceph is also on that single server. i'm still waiting for the replication to finish ... how can i stop it, i dont care about the current objects | 14:50 |
| bboris | okay then. i'll be here tonight i guess | 14:50 |
| bboris | got to figure that performance issue | 14:51 |
| creiht | bboris: start with trying to see what is bottlenecking during your perf run | 14:52 |
| creiht | either cpu is maxed on one of the servers | 14:52 |
| creiht | or disk io is maxed | 14:52 |
| creiht | once you figure that out, then we can help you tweak | 14:52 |
| creiht | gotta run, but will be back later | 14:52 |
| bboris | creiht: i will, thanks. later | 14:53 |
| *** foexle has quit IRC | 14:53 | |
| *** miqui has joined #openstack-swift | 14:59 | |
| *** mmcardle has quit IRC | 15:03 | |
| *** kevinc_ has joined #openstack-swift | 15:08 | |
| *** mmcardle has joined #openstack-swift | 15:16 | |
| *** nacim has quit IRC | 15:16 | |
| bboris | how do i stop the current replication? i don't care about the objects that swift is currently replicating and want to get rid of them so i can continue testing. | 15:19 |
| *** mmcardle has quit IRC | 15:28 | |
| *** mmcardle has joined #openstack-swift | 15:30 | |
| *** Anticimex has quit IRC | 15:34 | |
| *** Anticimex has joined #openstack-swift | 15:35 | |
| *** chandankumar has quit IRC | 15:44 | |
| nthacker_ | bboris, the way i have done it is used a single copy in my ring, but i am guessing you can also choose to not run the object-replicator by changing the conf file ( i havent tried this out though so cant comment on it) | 15:52 |
| *** stevemac has joined #openstack-swift | 15:55 | |
| *** mwstorer has joined #openstack-swift | 15:56 | |
| *** stevemac has quit IRC | 15:58 | |
| *** stevemac has joined #openstack-swift | 15:58 | |
| *** kevinc_ has quit IRC | 16:02 | |
| *** gyee has joined #openstack-swift | 16:03 | |
| stevemac | hi guys | 16:03 |
| *** yuanz has joined #openstack-swift | 16:04 | |
| *** kevinc_ has joined #openstack-swift | 16:07 | |
| *** yuan has quit IRC | 16:08 | |
| clayg | creiht: I think the default worker is auto now, doesn't saio use that? | 16:08 |
| clayg | bboris: just swift-init all stop (pkill -f swift if it comes to it) then rm -fr /srv/node/*/objects* | 16:09 |
| clayg | bboris: maybe service rsync stop in there somewhere | 16:10 |
| *** bboris has quit IRC | 16:11 | |
| *** bboris has joined #openstack-swift | 16:12 | |
| *** dANOKELOFF has quit IRC | 16:37 | |
| clayg | so is https://review.openstack.org/#/c/91133/4 solving like the *opposite* of the un-yielding starvation issue? | 16:38 |
| clayg | like on the one side, before we had the threadpool, write was blocking, and if your client was pushing bytes to the object server fairly fast you'd never trampoline on socket recv so that one request would starve everyone else (that's how it went right?) | 16:40 |
| *** tellesnobrega has joined #openstack-swift | 16:42 | |
| clayg | but now that we *have* a threadpool we're gunna trampoline into run_in_thread no matter how few bytes are in the buffer to write, so there's all this extra overhead context switching spent in the hub instead of parsing the chunked bytes coming out of the socket buffer and sending down epic writes into the threadpool | 16:42 |
| tellesnobrega | hi, im starting a devstack with swift but im getting an error when trying to create the swifttestusers | 16:42 |
| clayg | with this change, we let a single request do more work before kicking it back to the hub? | 16:42 |
| tellesnobrega | it says that the crendentials are note valid | 16:42 |
| clayg | tellesnobrega: try a different pitch? | 16:43 |
| * clayg hangs his head in shame at his terrible pun | 16:43 | |
| clayg | tellesnobrega: so like ./devstack/stack.sh is failing? | 16:43 |
| clayg | tellesnobrega: cause like when I setup devstack I just get those users for free, i always figured it was scripted; i barely understand keystoneclient | 16:44 |
| tellesnobrega | it finishes, but when i try to use sahara and register and image it fails | 16:44 |
| tellesnobrega | i was looking for the path of the certificates and found that they are not there | 16:45 |
| *** annegent_ has quit IRC | 16:45 | |
| notmyname | good morning | 16:46 |
| clayg | whoa | 16:46 |
| clayg | so but sahara is like a services vs. the swifttestusers guys which are just like... users | 16:47 |
| tellesnobrega | yes | 16:47 |
| tellesnobrega | on top of that i get some other errors, but i thought that this might be the root of all the others | 16:48 |
| clayg | and which certificate now? you're trying to find the signing certificate that keystone hands out to... swift? or sahara? | 16:48 |
| tellesnobrega | nova --os-password admin --os-username swiftusertest2 --os-tenant-name swifttenanttest2 x509-create-cert /home/stack/devstack/accrc/swifttenanttest2/swiftusertest2-pk.pem /home/stack/devstack/accrc/swifttenanttest2/swiftusertest2-cert.pem | 16:49 |
| tellesnobrega | i was looking fot this -cert.pem file | 16:49 |
| tellesnobrega | or pk.pem | 16:49 |
| tellesnobrega | it might be a glance problem as wel | 16:49 |
| tellesnobrega | im not sure | 16:49 |
| tellesnobrega | im actually brainstorming to find the problem | 16:50 |
| patchbot | tellesnobrega: You've given me 5 invalid commands within the last minute; I'm now ignoring you for 10 minutes. | 16:50 |
| clayg | notmyname: what is this guy doing!? | 16:50 |
| *** zhiyan is now known as zhiyan_ | 16:50 | |
| clayg | tellesnobrega: don't worry about patchbot he's a jerk | 16:50 |
| * notmyname goes to shoot patchbot in the head for now | 16:50 | |
| tellesnobrega | clayg: no problem | 16:50 |
| *** patchbot has quit IRC | 16:51 | |
| *** zhiyan_ is now known as zhiyan | 16:51 | |
| clayg | tellesnobrega: but i don't get why a failing nova command is a swift issue? does that command like work with the demo user? | 16:51 |
| tellesnobrega | yes | 16:51 |
| tellesnobrega | it fails only when creating swift users | 16:51 |
| clayg | what about swift stat --os-password --os-username swiftusertest2 --os-tenant-name swifttenanttest2 | 16:52 |
| clayg | umm... but that user is already created? | 16:52 |
| *** rook_ has joined #openstack-swift | 16:53 | |
| rook_ | anybody installed swift 1.13 via rpm packages on centos 6.4 ? (ducks) | 16:54 |
| clayg | tellesnobrega: i'm trying to read up on x590-create-cert but it seems to be some sort of eucatools shim - are you sure this workflow is correct for your usecase? I really don't think i know what you're trying to do. | 16:55 |
| tellesnobrega | im just trying to start a devstack installation with sahara | 16:56 |
| tellesnobrega | sahara uses swift, so i need it running, but maybe the problem isnt with swift itself | 16:56 |
| clayg | is #openstack-sahara a thing? | 16:57 |
| notmyname | clayg: I'm working on getting https://review.openstack.org/#/c/103731/ through jenkins this morning, then I'll land the 2 backport patches and we can cut an RC2 | 16:58 |
| clayg | notmyname: ok, zaitcev caught a typo in the commit - you want me to fix it up - or stay out the way of the merge train? | 17:01 |
| zaitcev | notmyname: commit message actually, not even code comments | 17:02 |
| *** shri has joined #openstack-swift | 17:04 | |
| clayg | i'm afk for a bit | 17:06 |
| *** zhiyan is now known as zhiyan_ | 17:08 | |
| acoles | notmyname: i may not make meeting today. no issues to report wrt SP | 17:15 |
| notmyname | acoles: thanks. I'll be getting an rc2 with the two small issues clayg found with the reconciler and container sync. it shouldn't invalidate any of your existing testing | 17:16 |
| peluse_ | notmyname: and my 2.0 update, going pretty good although I clearly don't spend enough time manually installing/operating swift... few more hiccups to work through and I'll be done w/rolling upgrades w/rsync will then do ssync real quick | 17:16 |
| tellesnobrega | clayg: yes, i'm there already | 17:20 |
| tellesnobrega | thanks anyways | 17:20 |
| bboris | does 1 replica mean a partition and 1 copy of it, or just a single partition? | 17:34 |
| notmyname | bboris: 1 replica means that you have the data in one and only one place | 17:35 |
| notmyname | 2 replicas == 2 locations, etc | 17:35 |
| *** mmcardle has quit IRC | 17:35 | |
| bboris | notmyname: thanks :) | 17:35 |
| *** bboris has quit IRC | 17:42 | |
| *** andyandy has quit IRC | 17:46 | |
| *** openstackgerrit has quit IRC | 17:49 | |
| *** mlipchuk has quit IRC | 17:49 | |
| *** openstackgerrit has joined #openstack-swift | 17:50 | |
| notmyname | clayg: I fixed the commit message typo since jenkins failed again | 17:51 |
| *** annegent_ has joined #openstack-swift | 17:54 | |
| *** annegent_ has quit IRC | 18:03 | |
| openstackgerrit | A change was merged to openstack/swift: SimpleClient http proxying https://review.openstack.org/102598 | 18:14 |
| *** Kbee has joined #openstack-swift | 18:16 | |
| Kbee | Hello, Q - when are accounts or account objects created .. I don't see a api for it.. but i see in code that PUT is annotated @public | 18:20 |
| peluse_ | notmyname: FYI all rolling ugprade tests w/rsync completed OK. Not being an operator I stumbled through some non SP related stuff but noted that I had to restart memcached on the proxy after upgrading as cached info data didn't have policy info it. I assume that's a normal part of one's process for ugprading (ro restart memcached). Onto ssync after some meetings | 18:29 |
| *** jamiehannaford has quit IRC | 18:31 | |
| Kbee | anyone ? - when are accounts or account objects created .. I don't see a api for it.. but i see in code that PUT is annotated @public | 18:32 |
| *** jamiehannaford has joined #openstack-swift | 18:32 | |
| clayg | torgomatic: did you see the py26 failure on gzip/ring fix? seems unrelated to your change - i think it may have been some other test changes I did recently? | 18:32 |
| torgomatic | clayg: no, I didn't... does py26 not support getargspec or some nonsense? | 18:33 |
| * torgomatic goes to look | 18:33 | |
| clayg | torgomatic: i tested on py26 and py3 and it seemed fine - i think the failure is unrelated, py26 tests passed on the change earlier | 18:34 |
| torgomatic | huh... it passed check but failed gate, so it can work | 18:34 |
| torgomatic | oh I see; let me go fix it real quick | 18:35 |
| clayg | you see it!? i don't follow | 18:40 |
| clayg | why would status be an instance of Timeout? | 18:41 |
| clayg | oh nm, i see it :P | 18:42 |
| clayg | well... :\ | 18:42 |
| torgomatic | I thought I saw it, but I don't | 18:43 |
| clayg | heh | 18:43 |
| *** tongli has joined #openstack-swift | 18:43 | |
| torgomatic | I thought it was a swob.Response and it should be .status_int instead, but it's an httplib thing, so it's probably someone's FakeConn being bad at life | 18:44 |
| notmyname | clayg: torgomatic: is this in https://review.openstack.org/#/c/103731/ ? | 18:44 |
| clayg | doh, if isinstance(Timeout(), Exception) <- not gunna work :P | 18:44 |
| notmyname | (just catchign up after a meeting) | 18:44 |
| torgomatic | notmyname: nope, it's http://logs.openstack.org/15/97615/2/gate/gate-swift-python26/957c8b0/nose_results.html | 18:44 |
| clayg | notmyname: wut? | 18:45 |
| torgomatic | clayg: where's that? | 18:45 |
| clayg | a few places in FakeConn | 18:45 |
| notmyname | ah https://review.openstack.org/#/c/97615/ | 18:46 |
| *** cutforth has joined #openstack-swift | 18:46 | |
| Kbee | notmyname, creiht , Q - when are accounts or account objects created .. I don't see a api for it.. but i see in code that PUT is annotated @public | 18:47 |
| notmyname | Kbee: so there's a couple of different way, but let me tell you about the common way | 18:48 |
| notmyname | Kbee: if swift sees a request to PUT a container in an account, and the token given is valid for that operation, and the account doesn't exist, then swift will automatically create the account | 18:49 |
| *** sombrafam has quit IRC | 18:49 | |
| notmyname | Kbee: if the token is valid and the account doesn't exist and it's a read operation for the account, swift will short-circuit and return a default empty response | 18:49 |
| clayg | notmyname: yeah that one is an ordering thing :\ | 18:50 |
| clayg | notmyname: i'll fix that one, torgomatic you got the other one? | 18:50 |
| torgomatic | clayg: the timeout thing? yeah, running tests now | 18:50 |
| notmyname | Kbee: there is a config value (default to off) that must be turned on before PUT/DELETE to an account will be accepted. so the other way is that you have a separate proxy server with that turned on and firewalled off to only accept connections eg from trusted sources like your provisioning system | 18:51 |
| notmyname | Kbee: (ok, I said I would just tell you one way, but that ^^ is the other way. now you know everything) | 18:51 |
| Kbee | notmyname, hmmm.. I tried this using keystone.. Suppose you auth using keystone and try to create a container using an account name different from ( AUTH_<tenant_id) , it throws 404.. so is it keystone middleware which doesnot allow accees to an account other than tenant_id ? | 18:53 |
| notmyname | Kbee: that sounds reasonable (but I'm not a keystone expert | 18:53 |
| *** angelastreeter has joined #openstack-swift | 18:53 | |
| *** Midnightmyth has joined #openstack-swift | 18:53 | |
| notmyname | Kbee: reasonable == doesn't surprise me | 18:53 |
| *** openfly has quit IRC | 18:54 | |
| Kbee | notmyname, :) .. got it .. | 18:54 |
| *** bboris has joined #openstack-swift | 18:55 | |
| notmyname | swift team meetng in a few minutes in #openstack-swift | 18:58 |
| *** elambert has joined #openstack-swift | 18:58 | |
| notmyname | err...in #openstack-meeting | 18:58 |
| Kbee | notmyname, i dug into keystoneauth middleware.. and yes.. they check whether account matches tenatn_id -> https://gist.github.com/keshavab/0d48163467d5ff0a2f54 | 18:59 |
| Kbee | notmyname, thanks | 18:59 |
| notmyname | Kbee: np | 19:00 |
| notmyname | meeting time | 19:00 |
| *** kenhui has quit IRC | 19:06 | |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Fix order dependent test in proxy.test_server https://review.openstack.org/104306 | 19:20 |
| *** jamiehannaford has quit IRC | 19:21 | |
| openstackgerrit | Samuel Merritt proposed a change to openstack/swift: Fix exception raising in FakeConn https://review.openstack.org/104319 | 19:26 |
| *** Kbee has quit IRC | 19:28 | |
| *** angelastreeter has quit IRC | 19:34 | |
| *** jasond has joined #openstack-swift | 19:38 | |
| jasond | i noticed the swift-temp-url script requires v1 | 19:38 |
| jasond | https://github.com/openstack/swift/blob/master/bin/swift-temp-url#L61 | 19:38 |
| jasond | should that be updated to accept v1 or v2? | 19:38 |
| mattoliverau | Cool, with that done, I'm going back to bed, see you all in a few hours :) | 19:38 |
| clayg | creiht: you know how your always railing on pipeline construction - have you looked at how some of the other os projects do it? | 19:39 |
| zaitcev | jasond: um there's no v2 | 19:39 |
| zaitcev | or x2 or n3 or anything else but v1 | 19:39 |
| *** cutforth has quit IRC | 19:40 | |
| jasond | zaitcev: ah, i guess i assumed "swift -V 2.0" meant swift v2, but i see it's the auth protocol. thanks | 19:40 |
| *** angelastreeter has joined #openstack-swift | 19:47 | |
| *** stevemac has quit IRC | 20:04 | |
| *** stevemac has joined #openstack-swift | 20:05 | |
| peluse_ | clayg: there? | 20:11 |
| creiht | clayg: not recently | 20:15 |
| peluse_ | creiht: thx, will gather a bit more data on the issue I'm still seeing w/rolling upgrades, probably not a real problem but want to double check a few things | 20:16 |
| *** stevemac has quit IRC | 20:18 | |
| peluse_ | notmyname: you there? | 20:28 |
| notmyname | peluse_: yes. what's up? | 20:29 |
| peluse_ | notmyname: ahh, jsut texted you :) | 20:29 |
| notmyname | peluse_: ya, I got that too :-) | 20:29 |
| notmyname | peluse_: is in here fine, or do we need a phone? | 20:29 |
| peluse_ | notmyname: so clayg's keyError fix is along the same lines as this but doesn't address what I see | 20:29 |
| notmyname | ok | 20:29 |
| peluse_ | here is fine | 20:30 |
| peluse_ | here's the scenario: | 20:30 |
| notmyname | ( torgomatic is listening too) | 20:30 |
| torgomatic | lies | 20:30 |
| peluse_ | if I'm doing IO, banging away, and I upgrade my proxy from 1.13.1 to v2 (claygs patch) the PUt in proxy obj server tries to ready the storage_policy key from container_info and gets a keyError | 20:31 |
| notmyname | the probability that torgomatic is listening is >0 | 20:31 |
| peluse_ | one easy patch is to try/except on keyError and define policy_index to None if it fails | 20:31 |
| notmyname | peluse_: proxy obj controller? | 20:31 |
| peluse_ | or, like I said, reset memchache so the info is cleared and you don't get the error but guess that's bad :) | 20:31 |
| peluse_ | notmyname: yes | 20:31 |
| peluse_ | right at the top of PUT | 20:32 |
| notmyname | against master, what file and line number are you looking at? | 20:33 |
| peluse_ | I'm looking at clayg patch but its policy_index = req.headers.get..... like 3 lines into PUT | 20:33 |
| notmyname | controllers.obj.py:460? | 20:33 |
| peluse_ | yup | 20:34 |
| notmyname | peluse_: should be container_info.get('storage_policy')? | 20:34 |
| torgomatic | if I only had a time machine... | 20:34 |
| notmyname | torgomatic: well, the good news is that it doesn't matter when you get one | 20:34 |
| peluse_ | yes | 20:34 |
| peluse_ | torgomatic: well, that I wasn't sure of so I was asking. You get an unhandled exception | 20:35 |
| torgomatic | memcache would get the *headers* jammed into it, not the result of _headers_to_blah, and then the same compatibility logic that saves us during rolling upgrades would also work with stale memcache data | 20:35 |
| torgomatic | but yeah, now that you see the bug it's obvious how it happens ;) | 20:36 |
| torgomatic | "obvious once pointed out" | 20:36 |
| peluse_ | torgomatic: so an unhandled exception doesn't puke that reqeust? | 20:36 |
| torgomatic | peluse_: oh, it'll screw things up all right; it should turn into a 500 | 20:36 |
| peluse_ | torgomatic: yeah, my cosbench run fails if I don't catch the exception | 20:36 |
| notmyname | ok, ya, that's my next question. unpatched, what does this do? cause a 5xx to the client? | 20:37 |
| peluse_ | sure looks that way.... | 20:37 |
| peluse_ | 6 places total this can happen (one in LE middleware, 5 in proxy controllers obj) | 20:39 |
| notmyname | peluse_: as in 6 places that need to be patched or this PUT is called in 6 places? | 20:39 |
| TaiSHi | Hi everyone | 20:39 |
| TaiSHi | I'm planning on a structure and I'm considering swift as an object storage, since I'll be managing a great amount of small files | 20:40 |
| peluse_ | notmyname: 6 places where we pull the policy_index out of container_info so 6 places to patch worst case | 20:40 |
| peluse_ | err, 'storage_policy' out of container_info | 20:41 |
| torgomatic | either that or we make container_info do something smarter | 20:42 |
| peluse_ | torgomatic: how do you make a dict do somethign smarter? | 20:42 |
| notmyname | peluse_: clayg is telling me why this isn't bad | 20:42 |
| peluse_ | oh, you mean the function :) | 20:42 |
| TaiSHi | Structure would be mostly webservers that need to have their data equal at all times, occasionally (and temporarily) new webservers might fire up. Also, webservers (at least fixed ones) would act as storage nodes | 20:42 |
| torgomatic | :) | 20:42 |
| TaiSHi | Can swift do this ? | 20:42 |
| torgomatic | peluse_: well, it's sort of hard; it's easy to chuck a value in for 'storage_policy', but what should that value be? | 20:43 |
| peluse_ | torgomatic: if it doesn't exist it can be None and will be handled properly later on | 20:44 |
| torgomatic | I guess if the container was ever seen by a pre-policy proxy (which is how it got in memcache) then it's policy 0 | 20:44 |
| peluse_ | yup, for a patch 0 or None would work | 20:44 |
| torgomatic | okay, that sounds reasonable then | 20:44 |
| * peluse_ awaits clayg's insight as to why this isn't bad | 20:44 | |
| notmyname | peluse_: nm. just walking through it with him | 20:45 |
| notmyname | peluse_: headers_to_container_info() defaults to SP 0 | 20:45 |
| notmyname | but if it's cached, it won't get added | 20:46 |
| notmyname | so a patch to container_info() would probably be best. and default to 0 seems right | 20:46 |
| peluse_ | but that doesn't fix the exception at line 460 (or the other places), thats downstream | 20:47 |
| torgomatic | notmyname: which is why we should get our time machine and start to always have stored the headers in memcache ;) | 20:47 |
| notmyname | :-) | 20:47 |
| peluse_ | right? (about the exception) | 20:47 |
| notmyname | TaiSHi: simply, yes. (but I'm focusing on the issue peluse_ has found that may require a backport in our current release QA first) | 20:48 |
| clayg | torgomatic: I think we can just pass the memcache info through a fill in defaults function before handing it out to the app | 20:48 |
| TaiSHi | notmyname: It is not urgent, but I'd like to get it out of the way | 20:48 |
| notmyname | peluse_: well, yes handling exceptions is good, but if we can guarantee that the header is always in the result, then the code gets simpler, right? | 20:49 |
| TaiSHi | Ping me whenever you want, I'll most likely be around here | 20:49 |
| torgomatic | clayg: proxy.base.Controller.container_info() seems like a good place to stick a call to setdefault | 20:49 |
| notmyname | peluse_: ie instead of patching those 6 locations, patch container_info (or whatever) once | 20:49 |
| clayg | peluse_: we could also throw out cache entries that don't have all the good bits we want and make that shit go out to the container servers as they find them | 20:49 |
| peluse_ | notmyname: ya, that's why I said "worst case" 6 places :) | 20:49 |
| notmyname | :-) | 20:49 |
| peluse_ | clayg: sounds intersting, don't know how to do that quickly on my end though. In fact you or torgomatic can probably nail this much faster than me although I'm glad to handle it of course | 20:50 |
| notmyname | peluse_: can you put together a patch and unit test for it? then between clayg torgomatic and me well find two reviewers to land it | 20:50 |
| peluse_ | heh, I'm game for anything - clayg/torgomatic? | 20:50 |
| * peluse_ starts to put something together... | 20:52 | |
| *** miqui has quit IRC | 20:53 | |
| peluse_ | looks like the container_info() patch will be simple | 20:53 |
| notmyname | peluse_: thanks for catching that. after you submit the patch, I'd love to hear how you found it | 20:56 |
| peluse_ | you got it | 20:57 |
| *** kevinc_ has quit IRC | 21:02 | |
| *** kevinc_ has joined #openstack-swift | 21:03 | |
| *** openfly has joined #openstack-swift | 21:04 | |
| *** tellesnobrega has quit IRC | 21:14 | |
| *** gyee has quit IRC | 21:18 | |
| *** annegent_ has joined #openstack-swift | 21:18 | |
| peluse_ | notmyname: propose to master then? | 21:19 |
| torgomatic | peluse_: sounds good to me | 21:22 |
| peluse_ | cool, letting one last test finish and will hit 'go' | 21:23 |
| *** tellesnobrega has joined #openstack-swift | 21:27 | |
| openstackgerrit | paul luse proposed a change to openstack/swift: Fix potential missing key error in container_info https://review.openstack.org/104350 | 21:30 |
| torgomatic | peluse_: fwiw, you can condense those 3 lines to simply "info.setdefault('storage_policy', None)" | 21:32 |
| torgomatic | or if you're really into code golf, "info.setdefault('storage_policy')" | 21:32 |
| peluse_ | torgomatic: sweet... needs it one other spot though... I pushed before doing my system level upgrade test and won't make that mistake again :) | 21:43 |
| clayg | torgomatic: peluse_: I'm not sure that's exactly the plce that needs fixing is it? | 22:00 |
| clayg | oh.... i guess i was just expecting it closer to the memcache get_info since that's the place where it might not be there | 22:00 |
| *** tongli has quit IRC | 22:01 | |
| clayg | but i guess that covers everything | 22:01 |
| *** bboris has quit IRC | 22:03 | |
| peluse_ | well, I tried there but that can return None so there are two callers that need it that wouldn't get it. I could be wrong, will udpate it in just a second, just passed my system test but need one more line of coverage in unit :) | 22:03 |
| mattoliverau | Morning.. again :) | 22:07 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Fix exception raising in FakeConn https://review.openstack.org/104319 | 22:10 |
| clayg | peluse_: are you making another change, or are we just waiting on gerrit? | 22:12 |
| *** angelastreeter has quit IRC | 22:13 | |
| clayg | peluse_: i take back what i said about it being in the wrong spot - it's in the right spot for exactly the reason you gave. | 22:13 |
| peluse_ | hmm, OK. let me post what I have and you can take a look | 22:13 |
| openstackgerrit | paul luse proposed a change to openstack/swift: Fix potential missing key error in container_info https://review.openstack.org/104350 | 22:14 |
| peluse_ | clayg: take a gander, its functional but I don't have an assertion for the change at line 992 in base.py yet | 22:15 |
| peluse_ | clayg: I have to leave to pick up my son or he'll be stranded. Be back online in an hour or so feel free to move it forard as you see fit, I'll check back when I get on again | 22:20 |
| clayg | peluse_: i see a good place to add the test for def container_info method on base controller? | 22:30 |
| clayg | peluse_: are you hacking or you want me to push something up? | 22:30 |
| *** annegent_ has quit IRC | 22:34 | |
| *** Midnightmyth has quit IRC | 22:36 | |
| clayg | peluse_: k, i'm going to take that as you're not on it just right this minute | 22:56 |
| clayg | torgomatic: you see the test changes i snuk in with that FakeConn fixup you did? | 22:57 |
| clayg | torgomatic: it's got your name on it, and i've had a rough track record lately, so - HEADS UP! | 22:57 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Fix potential missing key error in container_info https://review.openstack.org/104350 | 22:57 |
| torgomatic | clayg: yeah, I +2'ed it already; just waiting for someone without code in the mix to get in there and approve it | 22:58 |
| clayg | wooooo clayg the plagiarist | 22:58 |
| clayg | torgomatic: speaking of doesn't yet have code in the mix; I think peluse_'s container_info fix is ready to go -> https://review.openstack.org/104350 | 22:59 |
| clayg | probably be a nice treat when he gets back online to see it's ready already ;) | 22:59 |
| *** rook_ has quit IRC | 22:59 | |
| peluse_ | clayg: muchas gracias! | 23:06 |
| peluse_ | torgomatic: saw your comment too and, yup, that's why I used '0'... | 23:07 |
| *** annegent_ has joined #openstack-swift | 23:12 | |
| notmyname | peluse_: looking | 23:14 |
| *** Tyger has joined #openstack-swift | 23:20 | |
| Tyger | Good morning/afternoon/evening/night as appropriate for your local timezone. | 23:23 |
| *** annegent_ has quit IRC | 23:26 | |
| Tyger | Anyone besides torgomatic feel like taking a look at a simple cleanup patch 102401? | 23:26 |
| Tyger | https://review.openstack.org/#/c/102401/ | 23:27 |
| *** gyee has joined #openstack-swift | 23:33 | |
| peluse_ | Tyger: curious, what's the main reason for that change? | 23:36 |
| *** kevinc_ has quit IRC | 23:37 | |
| Tyger | peluse_: Mostly just a cleanup for consistency; I don't really have a strong opinion of if it is the right way to do it, but I was grepping the logs for requests originating at the object services (replication, etc) and assumed it would be "object-replicator" and the like. | 23:38 |
| Tyger | If I made that assumption, other people might as well. | 23:39 |
| *** kevinc_ has joined #openstack-swift | 23:39 | |
| peluse_ | Tyger: OK, wanted to make sure I was missing something somewhere :) I guess I wonder if changing something like this, sort of a "title" in the logs could impact other folks who may already have tools for log parsing that are going to have to change now and wonder "what the heck did this change for?" kinda thing. No idea if that's a real thing to be considered about or not though... | 23:42 |
| Tyger | peluse_: That thought occurred to me as well. But like you, I have no idea if it's a consideration either. | 23:44 |
| Tyger | I suspect people would be more interested in requests from the proxy server rather than requests from replication or container updates. | 23:44 |
| Tyger | If they are interested in the origin of the request at all, that is. | 23:45 |
| peluse_ | that's why it takes a few reviewers I guess :) | 23:49 |
| Tyger | peluse_: Feel free to add that as a comment so other reviewers might consider it. | 23:52 |
| peluse_ | yup, will do | 23:53 |
| notmyname | Tyger: when proposing changes to logs, two things to keep in mind: (1) logs are considered sacred like an API. that is, assume it's a public interface that is only changed carefully. and (2) leave the http://docs.openstack.org/developer/swift/logs.html docs better than you found them | 23:53 |
| notmyname | re #1, ie things can be added, but reordering is right out. changing values isn't forbidden, but it is only done carefully | 23:54 |
| *** ho has joined #openstack-swift | 23:54 | |
| *** kevinc_ has quit IRC | 23:54 | |
| Tyger | notmyname: Noted. Do you have any opinions on changing the user agent of the object server/replicator/etc? It would make ammending the "Storage Node Logs" section on "referer" easier when it can just say that all services use the name of the service and process ID as their user agent, rather than having to specify that object server uses obj-server, etc | 23:56 |
| *** kevinc_ has joined #openstack-swift | 23:57 | |
| Tyger | Given that only the proxy server is specifically mentioned at the moment, that does sort of confirm my feeling that people mostly care about requests from the proxy server. | 23:58 |
| notmyname | Tyger: while I think your proposed changes are probably better, I'd have to think a little more about that particular issue. I suspect it's ok to change that, but I'm not sure. I'd especially like to get input from gholt, acoles, and portante | 23:58 |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 23:59 |
| Tyger | notmyname: Sounds like a good idea to me. | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!