| *** yuanz has joined #openstack-swift | 00:04 | |
| *** peluse has quit IRC | 00:05 | |
| *** peluse has joined #openstack-swift | 00:06 | |
| *** yuan has quit IRC | 00:07 | |
| tdasilva | Hi, I'm testing object expirer feature and I noticed that after an object has "expired" doing a listing (GET) on the container still returns the object name. Is this a desirable? seems like a bug to me | 00:10 |
|---|---|---|
| tdasilva | sending a GET on the object itself returns 404 as expected | 00:10 |
| anticw | tdasilva: expected | 00:15 |
| anticw | tdasilva: if you GET and the object has expired you get a 404 | 00:15 |
| anticw | container and actually freeing are async and usually happen a little later | 00:16 |
| anticw | so long as the expirers are running | 00:16 |
| tdasilva | anticw: right, i understand that the object is not actually deleted when it is "expired", but only after the daemon is done running. | 00:18 |
| tdasilva | anticw: but it seems confusing that the container would still return a list container that object there | 00:18 |
| *** h6w has joined #openstack-swift | 00:20 | |
| h6w | Morning all. :-) | 00:20 |
| h6w | I'm unsure how to interpret this from swift-recon: "[replication_failure] low: 0, high: 3, avg: 1.2, total: 6, Failed: 0.0%, no_result: 0, reported: 5" | 00:21 |
| h6w | Is that saying that 3 replications failed or 0.0% failed? | 00:21 |
| h6w | Or that the failure failed? | 00:21 |
| h6w | Conversely, replcation_attempted says "total: 2", but I have 5 devices with 100% weight and 5 replicas. | 00:23 |
| *** matsuhashi has joined #openstack-swift | 00:24 | |
| *** krtaylor has joined #openstack-swift | 00:28 | |
| *** bill_az has quit IRC | 00:36 | |
| *** RockKuo has joined #openstack-swift | 00:42 | |
| openstackgerrit | A change was merged to openstack/swift: Use eventlet instead of threading for timeout https://review.openstack.org/85782 | 00:50 |
| openstackgerrit | A change was merged to openstack/swift: Fix logging issue when services stop on py26 https://review.openstack.org/86684 | 00:50 |
| *** csd has quit IRC | 01:01 | |
| openstackgerrit | paul luse proposed a change to openstack/swift: yAdd Storage Policy Support to Container Sync https://review.openstack.org/86469 | 01:15 |
| *** Guest____ has joined #openstack-swift | 01:27 | |
| *** ccarrizo has joined #openstack-swift | 01:31 | |
| h6w | notmyname: ^^ Help please? :-D | 01:33 |
| h6w | The deeper problem is that I'm mounting both proxies successfully, but they're giving me different results. :-( | 01:34 |
| h6w | Specifically, I can't see any files at all on my second proxy. | 01:34 |
| h6w | So I'm presuming that something's gone wrong with the replication. | 01:35 |
| *** nosnos has joined #openstack-swift | 01:51 | |
| *** saschpe has quit IRC | 02:00 | |
| *** saschpe has joined #openstack-swift | 02:02 | |
| *** lpabon has quit IRC | 02:09 | |
| *** zackf has joined #openstack-swift | 02:10 | |
| *** Guest____ has quit IRC | 02:12 | |
| *** gyee has quit IRC | 02:17 | |
| *** ccarrizo has quit IRC | 02:30 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Load constraints from cluster info dictionary https://review.openstack.org/86710 | 02:36 |
| openstackgerrit | paul luse proposed a change to openstack/swift: Add Storage Policy Support to Container Sync https://review.openstack.org/86469 | 02:46 |
| *** haomaiw__ has quit IRC | 02:57 | |
| *** haomaiwa_ has joined #openstack-swift | 03:00 | |
| *** haomaiwa_ has quit IRC | 03:00 | |
| *** haomaiwang has joined #openstack-swift | 03:00 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: In-process swift server for functional tests https://review.openstack.org/66108 | 03:05 |
| portante | notmyname: hopefully we can get these in-proc func tests submitted so that we can use this to help verify storage policies and raise the bar for reliability, avoiding problems like we have seen in the past. | 03:07 |
| openstackgerrit | A change was merged to openstack/swift: Add includes of referenced SAIO bin scripts https://review.openstack.org/86681 | 03:10 |
| *** matsuhashi has quit IRC | 03:14 | |
| *** madhuri has joined #openstack-swift | 03:15 | |
| *** nosnos has quit IRC | 03:20 | |
| *** nosnos has joined #openstack-swift | 03:20 | |
| *** nosnos has quit IRC | 03:25 | |
| *** matsuhashi has joined #openstack-swift | 03:28 | |
| openstackgerrit | Yuan Zhou proposed a change to openstack/swift: Clean up swift-object-info/swift-get-nodes bin files https://review.openstack.org/86593 | 03:42 |
| *** chandan_kumar has joined #openstack-swift | 03:48 | |
| *** SpamapS has left #openstack-swift | 03:49 | |
| *** piousbox has quit IRC | 03:49 | |
| *** MooingLemur has quit IRC | 04:02 | |
| *** nosnos has joined #openstack-swift | 04:14 | |
| openstackgerrit | Jenkins proposed a change to openstack/swift: Updated from global requirements https://review.openstack.org/81207 | 04:17 |
| *** MooingLemur has joined #openstack-swift | 04:21 | |
| *** chandan_kumar has quit IRC | 04:22 | |
| *** chandan_kumar has joined #openstack-swift | 04:35 | |
| *** MooingLemur has quit IRC | 04:36 | |
| *** MooingLemur has joined #openstack-swift | 04:44 | |
| *** praveenkumar has quit IRC | 04:46 | |
| *** praveenkumar has joined #openstack-swift | 04:48 | |
| *** RockKuo has quit IRC | 04:59 | |
| *** RockKuo has joined #openstack-swift | 05:00 | |
| *** Longgeek_ has joined #openstack-swift | 05:06 | |
| *** saju_m has joined #openstack-swift | 05:16 | |
| *** ppai has joined #openstack-swift | 05:23 | |
| *** chandan_kumar has quit IRC | 05:24 | |
| *** ashish_ has joined #openstack-swift | 05:29 | |
| *** Guest__ has joined #openstack-swift | 05:34 | |
| *** chandan_kumar has joined #openstack-swift | 05:37 | |
| *** zhiyan_ is now known as zhiyan | 05:39 | |
| *** zackf has quit IRC | 05:39 | |
| *** Guest__ has quit IRC | 05:47 | |
| *** Guest__ has joined #openstack-swift | 05:48 | |
| *** zaitcev has quit IRC | 06:01 | |
| *** RockKuo has quit IRC | 06:05 | |
| *** Guest__ has quit IRC | 06:06 | |
| *** Manish_ has joined #openstack-swift | 06:09 | |
| *** Guest__ has joined #openstack-swift | 06:10 | |
| *** psharma has joined #openstack-swift | 06:22 | |
| openstackgerrit | Madhuri Kumari proposed a change to openstack/swift: Print 'Container Count' in data base info https://review.openstack.org/86807 | 06:28 |
| *** MooingLemur has quit IRC | 06:32 | |
| *** chandan_kumar has quit IRC | 06:49 | |
| *** cheri has joined #openstack-swift | 06:51 | |
| *** chandan_kumar has joined #openstack-swift | 06:56 | |
| *** nshaikh has joined #openstack-swift | 06:57 | |
| *** haomaiwang has quit IRC | 07:14 | |
| *** haomaiwang has joined #openstack-swift | 07:14 | |
| *** cheri has quit IRC | 07:17 | |
| *** cheri has joined #openstack-swift | 07:18 | |
| *** PradeepChandani has joined #openstack-swift | 07:22 | |
| *** matsuhashi has quit IRC | 07:35 | |
| *** matsuhashi has joined #openstack-swift | 07:35 | |
| *** foexle has joined #openstack-swift | 07:40 | |
| *** Guest__ has quit IRC | 07:43 | |
| Manish_ | Hello Everyone | 07:45 |
| openstackgerrit | Yuan Zhou proposed a change to openstack/swift: Fix parsing storage policy in swift.conf https://review.openstack.org/86814 | 07:46 |
| Manish_ | does anyone knows if SWIFT has support for IPv6? | 07:46 |
| Manish_ | if yes,then since which release? | 07:46 |
| *** nacim has joined #openstack-swift | 07:47 | |
| *** foexle_ has joined #openstack-swift | 07:49 | |
| *** mmcardle has joined #openstack-swift | 07:54 | |
| *** ashish_ has quit IRC | 07:55 | |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Add headers to DirectClientException https://review.openstack.org/86817 | 07:57 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Add object-reconciler daemon https://review.openstack.org/85908 | 07:57 |
| *** mmcardle1 has joined #openstack-swift | 07:58 | |
| *** mmcardle has quit IRC | 07:58 | |
| *** MooingLemur has joined #openstack-swift | 08:15 | |
| *** matsuhas_ has joined #openstack-swift | 08:25 | |
| *** matsuhashi has quit IRC | 08:25 | |
| *** jamie_h has joined #openstack-swift | 08:35 | |
| *** joeljwright has joined #openstack-swift | 08:38 | |
| redbo | Manish_: I don't know which release, but swift has supported ipv6 for over 3 years. I'm not sure if anyone's actually running a cluster all on ipv6, though. | 08:39 |
| openstackgerrit | Yuan Zhou proposed a change to openstack/swift: Fix parsing storage policy in swift.conf https://review.openstack.org/86814 | 08:40 |
| *** d89 has joined #openstack-swift | 08:42 | |
| Manish_ | redbo: thanks a lot | 08:43 |
| *** mkerrin has quit IRC | 08:46 | |
| *** haomaiw__ has joined #openstack-swift | 08:48 | |
| *** haomaiwang has quit IRC | 08:52 | |
| *** mkerrin has joined #openstack-swift | 08:52 | |
| openstackgerrit | Yuan Zhou proposed a change to openstack/swift: Adding deprecate policy support https://review.openstack.org/86827 | 08:52 |
| *** RockKuo has joined #openstack-swift | 08:58 | |
| *** tanee-away is now known as tanee | 08:59 | |
| *** _bluev has joined #openstack-swift | 09:03 | |
| *** chandan_kumar has quit IRC | 09:10 | |
| *** chandan_kumar has joined #openstack-swift | 09:15 | |
| *** ccorrigan has joined #openstack-swift | 09:16 | |
| openstackgerrit | Takashi Kajinami proposed a change to openstack/swift: Add timestamp checking in AccountBroker.is_status_deleted https://review.openstack.org/86248 | 09:24 |
| *** saschpe has quit IRC | 09:27 | |
| *** saschpe has joined #openstack-swift | 09:29 | |
| openstackgerrit | Takashi Kajinami proposed a change to openstack/swift: Add timestamp checking in AccountBroker.is_status_deleted https://review.openstack.org/86248 | 09:33 |
| grapsus_ | hi there | 09:37 |
| *** saju_m has quit IRC | 09:45 | |
| _bluev | Hello | 09:45 |
| *** chandan_kumar has quit IRC | 09:47 | |
| *** zhiyan is now known as zhiyan_ | 09:55 | |
| *** saju_m has joined #openstack-swift | 10:01 | |
| *** nosnos has quit IRC | 10:03 | |
| *** matsuhas_ has quit IRC | 10:04 | |
| *** foexle_ has quit IRC | 10:21 | |
| *** foexle has quit IRC | 10:21 | |
| *** Anju has quit IRC | 10:23 | |
| *** saju_m has quit IRC | 10:25 | |
| *** foexle has joined #openstack-swift | 10:26 | |
| *** foexle_ has joined #openstack-swift | 10:27 | |
| *** chandan_kumar has joined #openstack-swift | 10:29 | |
| openstackgerrit | Yuan Zhou proposed a change to openstack/swift: Fix parsing storage policy in swift.conf https://review.openstack.org/86814 | 10:31 |
| *** tdasilva has left #openstack-swift | 10:39 | |
| *** fifieldt has quit IRC | 10:47 | |
| *** Manish_ has quit IRC | 10:52 | |
| *** ccorrigan has quit IRC | 10:57 | |
| *** ccorrigan has joined #openstack-swift | 10:58 | |
| *** Trixboxer has joined #openstack-swift | 11:14 | |
| *** judd7_ has quit IRC | 11:15 | |
| *** RockKuo has quit IRC | 11:19 | |
| *** Longgeek_ has quit IRC | 11:19 | |
| *** jamie_h has quit IRC | 11:30 | |
| *** jamie_h has joined #openstack-swift | 11:32 | |
| *** cheri has quit IRC | 11:32 | |
| *** chandan_kumar has quit IRC | 11:41 | |
| *** mkollaro has joined #openstack-swift | 11:43 | |
| *** ppai has quit IRC | 11:50 | |
| *** chandan_kumar has joined #openstack-swift | 11:54 | |
| *** psharma has quit IRC | 12:00 | |
| openstackgerrit | paul luse proposed a change to openstack/swift: Add Storage Policy Support to Container Sync https://review.openstack.org/86469 | 12:20 |
| *** anticw_ has joined #openstack-swift | 12:24 | |
| *** rahmu has quit IRC | 12:24 | |
| *** anticw has quit IRC | 12:24 | |
| *** hurricanerix_ has quit IRC | 12:24 | |
| *** rahmu has joined #openstack-swift | 12:24 | |
| *** hurricanerix has joined #openstack-swift | 12:24 | |
| *** mmcardle1 has quit IRC | 12:31 | |
| *** MooingLemur has quit IRC | 12:35 | |
| *** judd7 has joined #openstack-swift | 12:42 | |
| *** mmcardle has joined #openstack-swift | 12:56 | |
| *** praveenkumar has quit IRC | 13:15 | |
| *** jamie_h has quit IRC | 13:24 | |
| *** jamie_h has joined #openstack-swift | 13:25 | |
| *** lpabon has joined #openstack-swift | 13:31 | |
| *** NM1 has joined #openstack-swift | 13:42 | |
| *** praveenkumar has joined #openstack-swift | 13:54 | |
| *** nacim has quit IRC | 14:03 | |
| *** nacim has joined #openstack-swift | 14:04 | |
| *** russellb is now known as rustlebee | 14:07 | |
| *** zhiyan_ is now known as zhiyan | 14:09 | |
| *** mwstorer has joined #openstack-swift | 14:09 | |
| *** nshaikh has quit IRC | 14:11 | |
| *** tdasilva has joined #openstack-swift | 14:13 | |
| *** zackf has joined #openstack-swift | 14:15 | |
| *** piyush1 has joined #openstack-swift | 14:25 | |
| *** chandan_kumar has quit IRC | 14:31 | |
| notmyname | FYI, tomorrow morning (Paris time) when ttx wakes up, he'll cut an RC2 for swift, assuming we haven't found anything else today. | 14:44 |
| notmyname | if something comes up today that needs to be backported and we get it landed on milestone-proposed, it can be included in rc2 | 14:44 |
| portante | notmyname: great, thanks | 14:44 |
| briancline | is that 1.13.1? | 14:45 |
| notmyname | briancline: yes. we've currently got 1.13.1-rc1 as our RC for icehouse. | 14:45 |
| * briancline has been under several rocks/glaciers for a while | 14:46 | |
| notmyname | briancline: 1.13.1 final will be included in the openstack icehouse release | 14:46 |
| creiht | notmyname: cool | 14:46 |
| briancline | excellent | 14:46 |
| notmyname | and the icehouse release will be cut net thursday the 17th | 14:46 |
| briancline | man, almost halfway through april... guess it's time to start putting summit slides together | 14:47 |
| notmyname | creiht: so if anything suspicious pops up today, let everyone know sooner than later :-) | 14:47 |
| notmyname | briancline: how's working for Big Corp going? | 14:48 |
| notmyname | did you get your IBM songbook? | 14:48 |
| briancline | oh yes, it came in the briefcase with white dress shirt, black pants, and black tie | 14:49 |
| briancline | ;) | 14:49 |
| creiht | notmyname: will do | 14:50 |
| notmyname | nice | 14:50 |
| notmyname | briancline: will you be going to atlanta? | 14:50 |
| creiht | lol | 14:50 |
| notmyname | creiht: are you not at pycon? | 14:50 |
| creiht | notmyname: not this year | 14:50 |
| notmyname | creiht: ya, me neither :-( | 14:51 |
| creiht | I'll probably go next year though | 14:51 |
| notmyname | are they doing 2 years in canada? | 14:51 |
| creiht | yeah | 14:52 |
| creiht | notmyname: I think I am going to switch to going to every other pycon | 14:53 |
| briancline | notmyname: in all seriousness it's been going pretty well, though. working on some fun object storage things lately, hopefully some of that time will clear up soon to allow us to get back to co-championing the search blueprint | 14:53 |
| notmyname | cool | 14:53 |
| briancline | notmyname: and yup, I'll definitely be there all week. stayed through end of Friday due to the ops sessions they're doing now. seems intriguing | 14:54 |
| briancline | s/stayed/will be staying/ | 14:54 |
| briancline | thought that was a nice addition to the format | 14:55 |
| notmyname | there are several new things this year. it will be interesting to see how it works out | 14:55 |
| briancline | nice. I'm excited to see the place, whilst booking I realized I stayed at a nearby Sheraton sometime in college during a visit, but didn't get to make it out to the park/cnn center/etc. | 14:57 |
| notmyname | I'm told that the tentative summit schedule is public and I can reference it. https://docs.google.com/spreadsheet/ccc?key=0AmUn0hzC1InKdGNXcWlWX0FIekQxbUtvRVlnVF9IV3c&usp=drive_web#gid=5 | 14:57 |
| ctennis | notmyname: there's some discussion above about swift working with ipv6..is that true? I thought there was a problem with eventlet that prevented it. | 14:57 |
| notmyname | ctennis: redbo said it worked :-) | 14:58 |
| creiht | lol | 14:58 |
| creiht | ctennis: there was a lot of work done a while ago to make sure that it worked | 14:58 |
| notmyname | ctennis: I know there were some patches to make it work a long time ago, but like redbo said I don't know anyone who is using it who can confirm (including myself) | 14:58 |
| ctennis | interesting, I guess I need to look into it | 14:58 |
| creiht | at that time at least | 14:58 |
| *** csd has joined #openstack-swift | 14:58 | |
| ctennis | I didn't know you could set the node addresses in the ring as ipv6 | 14:58 |
| notmyname | ctennis: so I'd consider it a bug if it doesn't work. (it's like we need tests for that...) | 14:59 |
| ctennis | ok good to know. My main interest here is supporting encrypted network traffic between nodes, which is a bit easier with ipsec and ipv6. | 14:59 |
| notmyname | cool! | 15:00 |
| * notmyname asks ctennis in a different channel who's doing that... | 15:00 | |
| creiht | I think there was one bug that popped up where getmyips wouldn't work for ipv6 addrs | 15:00 |
| creiht | because a library change happened or something like that | 15:00 |
| *** Gu_______ has joined #openstack-swift | 15:01 | |
| * notmyname goes offline for a bit to get ready for the day | 15:03 | |
| briancline | speaking of tests, am I the only one with beef about TestAccount.test_bad_metadata? | 15:04 |
| joeljwright | Is there a problem with the test infrastructure? The current zuul job stats graphs don't look good http://status.openstack.org/zuul/ and some jobs have been waiting for hours to run | 15:04 |
| briancline | joeljwright: not sure, there was a huge uptick in jobs last night that probably caused the backlog. I've got 2 for glance that have been painstakingly slow | 15:05 |
| joeljwright | I just noticed the zuul job queue runnin/worker numbers drop towards 0 so I thought it was worth mentioning | 15:06 |
| joeljwright | waiting/total jobs are now both near to 1500 | 15:07 |
| joeljwright | also, zuul jobs launched per hour appears to be 0 :S | 15:07 |
| briancline | also I may not be reading it correctly, but the total nodes graph doesn't seem to roughly match up numbers-wise with the job queue workers graph | 15:07 |
| briancline | oh wow, yeah, didn't notice that drop in workers | 15:07 |
| briancline | seems like a lot of the delays are due to python3.3 checks | 15:12 |
| joeljwright | I can only find 5 running tests atm | 15:12 |
| joeljwright | all in check-tripleo queue | 15:12 |
| joeljwright | everything else is stuck in queued | 15:13 |
| *** tanee is now known as tanee-away | 15:13 | |
| briancline | joeljwright: here you go, 0411 1018 < fungi> mrmartin: trouble adding new workers to run jobs. i'm currently trying to sort it out | 15:19 |
| *** tanee-away is now known as tanee | 15:20 | |
| *** Gu_______ has quit IRC | 15:20 | |
| joeljwright | briancline: thanks for checking | 15:21 |
| joeljwright | hope it's not a major issue! | 15:22 |
| *** foexle_ has quit IRC | 15:23 | |
| *** zul has quit IRC | 15:26 | |
| *** foexle has quit IRC | 15:26 | |
| *** csd has quit IRC | 15:29 | |
| *** csd has joined #openstack-swift | 15:29 | |
| *** zul has joined #openstack-swift | 15:30 | |
| tdasilva | hi, i'd like to take a stab and fixing bug: 1076202 and while testing object-expirer, I found a couple of issues...wondering if anyone can offer some help. | 15:37 |
| tdasilva | first, I put two objects, one set to expire and the other not. sending a GET request on the expired object resulted in a 404 as expected, but now when running the object-expirer daemon it logs "0 objects expired", any ideas? | 15:37 |
| *** davidhadas_ has joined #openstack-swift | 15:43 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Load constraints from cluster info dictionary https://review.openstack.org/86710 | 15:44 |
| *** G________ has joined #openstack-swift | 15:46 | |
| *** davidhadas has quit IRC | 15:46 | |
| *** dmsimard has quit IRC | 15:51 | |
| anticw_ | tdasilva: container (and account) listings aren't always up-to-date | 15:58 |
| anticw_ | even w/o expirey it's pretty easy to get the container server days behind right now | 15:59 |
| *** judd7 has quit IRC | 16:01 | |
| tdasilva | well..it's not so much the listing of the container, it's more the fact that the expired object is not getting picked up by the expirer daemon to be picked up | 16:01 |
| tdasilva | anticw_: this call self.swift.get_account_info(self.expiring_objects_account) | 16:02 |
| tdasilva | is returning 0, 0 | 16:02 |
| tdasilva | oh wait, you said "days behind" | 16:04 |
| tdasilva | so an object could be expired for days, but because the container server is behind it will never get picked up by the expirer daemon? | 16:06 |
| *** gyee has joined #openstack-swift | 16:07 | |
| *** nacim has quit IRC | 16:12 | |
| notmyname | pandemicsyn: irish brownies this morning https://twitter.com/notmyname/status/454653654172045312 | 16:14 |
| notmyname | err...cupcakes. not brownies | 16:16 |
| *** G________ has quit IRC | 16:31 | |
| *** foexle has joined #openstack-swift | 16:39 | |
| *** zhiyan is now known as zhiyan_ | 16:40 | |
| *** zhiyan_ is now known as zhiyan | 16:40 | |
| *** zhiyan is now known as zhiyan_ | 16:43 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Load constraints from cluster info dictionary https://review.openstack.org/86710 | 16:45 |
| *** praveenkumar has quit IRC | 16:47 | |
| *** G________ has joined #openstack-swift | 16:49 | |
| *** foexle has quit IRC | 16:50 | |
| *** piyush1 has quit IRC | 16:52 | |
| *** zaitcev has joined #openstack-swift | 17:05 | |
| *** ChanServ sets mode: +v zaitcev | 17:05 | |
| *** G________ has quit IRC | 17:07 | |
| *** _bluev has quit IRC | 17:13 | |
| *** d89 has quit IRC | 17:17 | |
| *** tanee is now known as tanee-away | 17:17 | |
| pandemicsyn | notmyname: that looks/sounds damn tasty! | 17:29 |
| notmyname | pandemicsyn: oh, it was | 17:29 |
| *** chandan_kumar has joined #openstack-swift | 17:35 | |
| tdasilva | notmyname: i'm trying to test object-expirer, but getting conflicting results. is there anywhere you can point me that describes the expected behavior | 17:37 |
| tdasilva | ? | 17:37 |
| notmyname | tdasilva: http://docs.openstack.org/developer/swift/overview_expiring_objects.html is where I'd start, but you've probably already read it | 17:38 |
| tdasilva | yep | 17:38 |
| *** piyush1 has joined #openstack-swift | 17:39 | |
| tdasilva | notmyname: i guess my question is: is it correct to expect that once an object is expired, the next time the expirer daemon runs, that object should be deleted? | 17:39 |
| tdasilva | notmyname: anticw_ answers seems to point to an issue with account and container listing being behind for days, so that's why it might take some time for the daemon to find the expired object | 17:40 |
| *** mmcardle has quit IRC | 17:40 | |
| *** piyush2 has joined #openstack-swift | 17:40 | |
| tdasilva | that seems strange, so I just wanted to confirm | 17:41 |
| *** piyush1 has quit IRC | 17:43 | |
| zaitcev | It does seem strange on the surface of it. The expirer works through proxy, does it not? Therefore, any deletions it does should be properly reflected in container and account listings through normal means. If Chris' cluster gets divergent listings that persist for "days", then his cluster is broken. However, I do not have first-hand experience with large clusters, and it's possible that we have some kind of bug that preven | 17:45 |
| zaitcev | ts normal updates to propagate when expirer works. | 17:45 |
| notmyname | tdasilva: teh expirer uses an internal swift account to track what needs to be deleted. yes, once the object is expired, it should be deleted the next time the expirer runs. and thus, indirectly, the listing will be updated | 17:46 |
| notmyname | tdasilva: zaitcev: listings being out of sync for days is an issue, but unfortunately not an uncommon one. as containers get large (and busy), things have trouble staying in sync sometimes. gholt would have all the war stories for you :-) | 17:47 |
| *** piousbox has joined #openstack-swift | 17:48 | |
| tdasilva | notmyname, zaitcev: I started with a very simple cluster setup of one node and no replication. I put two objects there, one with "expiration" header | 17:48 |
| tdasilva | that's why i thought that it would be very strange to not see the listing be updated right away | 17:49 |
| zaitcev | indeed | 17:49 |
| zaitcev | well, unless you forgot to start updaters or something | 17:49 |
| zaitcev | "no replication" does not mean that you can get away with not running "swift-init start rest" duh | 17:50 |
| tdasilva | zaitcev: that might be it, in fact i did not start the updater | 17:52 |
| tdasilva | zaitcev: would the expirer daemon depend on that too? because it just uses the InternalClient to get a list of container and objects in the "hidden" .expiring-objects account | 17:53 |
| zaitcev | tdasilva: I thought Proxy just stored directly into that hidden account if it detected an expiration header, but truthfuly I can never remember how it works. My mind is too small. | 17:55 |
| openstackgerrit | paul luse proposed a change to openstack/swift: Add Storage Policy Documentation https://review.openstack.org/85824 | 17:57 |
| *** foexle has joined #openstack-swift | 18:00 | |
| *** ashish_ has joined #openstack-swift | 18:05 | |
| clayg | peluse: yeah the container-obj-put-409 isn't strictly needed if the cleanup table approach ends up looking better in the end | 18:12 |
| clayg | I'll probably rebase container-obj-put-409 on the reconciler change when I add the object-updater enqueue bits - so anything you can pull out of that patch is good | 18:12 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: In-process swift server for functional tests https://review.openstack.org/66108 | 18:14 |
| portante | zaitcev: when you have a chance, can you look at the /info use in func tests, https://review.openstack.org/86710 and then the in-process functional tests, https://review.openstack.org/66108 ? | 18:17 |
| zaitcev | portante: sure | 18:17 |
| portante | I believe all of the concerns have been addressed. | 18:17 |
| zaitcev | just going to re-push https://review.openstack.org/85909 real quick, Gerrit says it can't merge (due to get_log_line in import lists) | 18:18 |
| portante | zaitcev: okay | 18:18 |
| portante | clayg: the latest in-process code uses your work to automate the pipeline setup a bit more, thank you! | 18:19 |
| *** Guest__ has joined #openstack-swift | 18:20 | |
| *** Guest__ has quit IRC | 18:22 | |
| openstackgerrit | Pete Zaitcev proposed a change to openstack/swift: Separate Backend from Broker https://review.openstack.org/85909 | 18:23 |
| *** Guest__ has joined #openstack-swift | 18:23 | |
| clayg | portante: oh neato, I wasn't sure I got the logging quite duped up quite right - but I'd acctually prefer the simpler maintance burdon for pipeline changes vs. pristine clear syslog during in process functests. | 18:24 |
| zaitcev | portante: Don't you think that Clay got a point about retaining a possibility to override constraints? | 18:26 |
| zaitcev | Not that I ever used them | 18:27 |
| portante | if a cluster offers /info, why would you want to override the constraints? | 18:28 |
| portante | what do you gain doing that? | 18:28 |
| portante | the functional tests are trying to ensure that the cluster does what it advertises | 18:28 |
| portante | in the absence of /info, somebody has to tell it, which is what test.conf is for | 18:29 |
| portante | if there is a good reason to have a set of tests use a different set of constraints from a running cluster, that would be nice to know | 18:29 |
| *** jasondotstar has joined #openstack-swift | 18:30 | |
| *** chandan_kumar has quit IRC | 18:30 | |
| zaitcev | The reason may be a case when you try to point out that RGW or other independent implementation is degraded. So, you run functests against RAX CF, then use same test.conf, change auth_host, run it against the cluster in question. | 18:31 |
| zaitcev | Using .functests as API tests in absence of a real compliance suite. | 18:32 |
| portante | zaitcev: you mean you would have potentially different constraint values? or are you talking about changing auth_host? | 18:32 |
| zaitcev | I do not insist it matters, and you need to take it up with Clay, but since you asked... | 18:32 |
| zaitcev | No, SAME constraint values DIFFERENT auth_host. | 18:33 |
| zaitcev | Ouch, sorry. I get mad easily | 18:33 |
| zaitcev | pheeew | 18:34 |
| portante | ah, so then, if you look at these changes, we *always* read test.conf for the auth_host now, we only read test.conf "swift" section for constraints when /info dose not report back successfully | 18:34 |
| zaitcev | well, sure | 18:34 |
| zaitcev | but if /info says that your object name length is 5, is that a functional Swift cluster or a joke? | 18:35 |
| portante | what does it matter? | 18:35 |
| portante | the functional tests should try to verify that config | 18:36 |
| zaitcev | okay | 18:36 |
| portante | one more update to the in-process, missing container_sync middleware from the pipeline | 18:37 |
| portante | for in-process, that is | 18:37 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: In-process swift server for functional tests https://review.openstack.org/66108 | 18:37 |
| clayg | peluse: yuanz: I think deprecating policies would be a killer feature, but I don't think it can remove them from the POLICIES data structure - it needs to just stop lising them in /info and letting you create new containers for them - is there a trello card for this? | 18:38 |
| zaitcev | Jeezus, being on core for this project is no sinecure. | 18:38 |
| tdasilva | zaitcev: i'm running some more tests but I think you were right about the updaters, starting that seems to have done the trick | 18:39 |
| tdasilva | thanks | 18:39 |
| zaitcev | tdasilva: I already gave myself a few candied almonds for this, glad it worked. | 18:40 |
| tdasilva | lol | 18:40 |
| *** mmcardle has joined #openstack-swift | 18:41 | |
| peluse | clayg: there is a trello card talking about the topic but no specific actions identified, I like your take on it though. I'll update the card and see if Yuan can finish it off | 18:42 |
| clayg | peluse: cool! | 18:42 |
| peluse | clayg: container_sync is ready for your review. Got a "looks OK" from gholt and I figured out the Jenkins failure. Had to patch the rings after patching policies. Please have a look https://review.openstack.org/#/c/86469/ | 18:43 |
| clayg | peluse: will do | 18:44 |
| *** mmcardle has quit IRC | 18:45 | |
| * peluse wonders if his class patching was OK on that one | 18:45 | |
| *** Guest__ has quit IRC | 18:50 | |
| ashish_ | Hey everyone I am adding adding a new disk to swift.I have a doubt.I want to know in make-account-ring.sh do I need to add only the newly created account rings or both the existing rings and the new rings | 18:53 |
| *** chandan_kumar has joined #openstack-swift | 18:53 | |
| *** jamie_h has quit IRC | 18:54 | |
| *** Guest__ has joined #openstack-swift | 18:57 | |
| *** foexle has quit IRC | 18:58 | |
| zaitcev | toot toot [zaitcev@guren swift-fetch]$ SWIFT_TEST_IN_PROCESS=1 ./.functests | 19:05 |
| peluse | hmmm.... not sure what to make of that :) | 19:06 |
| zaitcev | ashish_: if you add, only add new ones. that way you preserve your old data | 19:06 |
| *** joeljwright has quit IRC | 19:07 | |
| ashish_ | zaitcev thanks.http://docs.openstack.org/developer/swift/admin_guide.html I did not understand the step 3 of scripting the ring creation.Could you hel | 19:08 |
| ashish_ | zaitcev thanks.http://docs.openstack.org/developer/swift/admin_guide.html I did not understand the step 3 of scripting the ring creation.Could you help | 19:08 |
| ashish_ | zaitcev thanks.I am following the following doc http://docs.openstack.org/developer/swift/admin_guide.html I did not understand the step 3 of scripting the ring creation.Could you help | 19:09 |
| zaitcev | portante: gtg, sorry - I'll re-review 66108 in a few hours | 19:11 |
| portante | zaitcev: no prob, thx | 19:12 |
| openstackgerrit | paul luse proposed a change to openstack/swift: Minor refactor of storage policy helper functions https://review.openstack.org/87001 | 19:15 |
| peluse | clayg: if that ^ looks good, I'll rebase 409 (as it needs it anyway) and I believe that will pull those changes out of the patch as, well, changes | 19:15 |
| peluse | ashish_: I might be able to help, looks like zaitcez might be busy. Which part is not clear? | 19:16 |
| *** joeljwright has joined #openstack-swift | 19:17 | |
| ashish_ | peluse thanks.When adding new disks to I need to set the rsyncd.conf again. | 19:18 |
| peluse | ashish_: one sec - phone call | 19:20 |
| peluse | ashish_: OK, I'm back. I don't believe so. Are you just setting up an all-in-one per http://docs.openstack.org/developer/swift/development_saio.html or is this editing something existing? | 19:23 |
| *** ashish_ has quit IRC | 19:23 | |
| *** joeljwright has quit IRC | 19:26 | |
| peluse | ashish_: guess it doesn't matter either way but the answer would be no since you should have configured rsync based on directories (/srv/node) with links below that to each device | 19:27 |
| *** simpleAJ has joined #openstack-swift | 19:35 | |
| simpleAJ | Hi, | 19:36 |
| simpleAJ | i was wondering if it is possible to query object store itself..directly like what proxy server is doing.. is there any test cases that I can play with | 19:36 |
| *** ashish_ has joined #openstack-swift | 19:36 | |
| clayg | simpleAJ: you can look at direct client, probetests use them some I believe - also the list_endpoints middleware may be useful... | 19:38 |
| tdasilva | clayg, simpleAJ: maybe the InternalClient might also be another place to look | 19:39 |
| *** mmcardle has joined #openstack-swift | 19:41 | |
| clayg | tdasilva: maybe, it doesn't really make the backend api's any more clear - but if I mis understood his question that could be exactly what he wants... | 19:42 |
| simpleAJ | yeah .. let me take a look at them | 19:43 |
| simpleAJ | ./unit/common/middleware/test_list_endpoints.py | 19:44 |
| simpleAJ | ./unit/common/test_internal_client.py | 19:44 |
| simpleAJ | I could find these...but I could not find "direct client" | 19:44 |
| simpleAJ | where should I look for them | 19:44 |
| simpleAJ | got it ./unit/common/test_direct_client.py | 19:44 |
| simpleAJ | :D | 19:44 |
| simpleAJ | thanks | 19:44 |
| clayg | peluse: I had a couple of comments on container sync, I think you need to get that policy index passed through on object get (and some tests on that for non-legacy policies) - is that the only backend request that container-sync makes? | 19:45 |
| *** mmcardle has quit IRC | 19:46 | |
| clayg | portante: zaitcev: fwiw i'm un-swayed, I think test.conf should be the higest priority, there was a comment that functests are for testing a swift cluster does what it reports and like... that can't have always been true because swift clusters didn't always have /info :D | 19:47 |
| *** csd has quit IRC | 19:47 | |
| *** mkollaro has quit IRC | 19:49 | |
| clayg | yuanz: notmyname: peluse: can we get https://review.openstack.org/#/c/86593/ listed on the pre-storage-policy topic or just stright up over to feature/ec as a dependency for https://review.openstack.org/#/c/82734/ | 19:52 |
| *** mmcardle has joined #openstack-swift | 19:56 | |
| portante | clayg: so if your test.conf constraints section does not match the values that the swift cluster is running with, how can those tests be valid? | 19:57 |
| portante | I just don't understand the case | 19:57 |
| portante | but maybe there is a good reason for it that I am not seeing | 19:57 |
| *** simpleAJ has left #openstack-swift | 19:58 | |
| *** mmcardle has quit IRC | 20:00 | |
| clayg | portante: usecases - 1) I just want to see the test turn red (verify it fails when the value doesn't match the cluster and give a good assertion error/failure) | 20:00 |
| ashish_ | clayg portante .I have a doubt. When adding new disks to I need to set the rsyncd.conf again. | 20:01 |
| clayg | 2) I want to make sure the cluster is enforcing the constraints it reports - many of the explicit constraint tests have negative tests baked in - but if a cluster does pass with a value other than reported it may be a bug in the test or deploy | 20:01 |
| clayg | ashish_: some of that is a property of how saio's are just weird wrt to rsync and vm_test_mode | 20:01 |
| clayg | ashish_: vagrant-swift-all-in-one does things quite a bit different, but in that config for example adding more disks (it's configurable at build time) doesn't require a different rsync.conf - most production systems expose all disks as well w/o any need to change rsync.conf when adding new harddrives | 20:02 |
| *** Guest__ has quit IRC | 20:03 | |
| *** Guest__ has joined #openstack-swift | 20:03 | |
| *** jasondotstar has quit IRC | 20:04 | |
| ashish_ | clayg thanks. I am following the doc http://docs.openstack.org/developer/swift/admin_guide.html to add a new loopback device to swift.Would this be sufficient. | 20:05 |
| portante | clayg: #1 and #2 sound an awful lot like the same thing, and shouldn't #2 be part of the functional tests themselves, as unless you really know what you are doing as a developer, it does not seem clear you can arrive at a conclusive indicator? | 20:06 |
| portante | but I'll flip hierarchy | 20:07 |
| clayg | portante: cool! | 20:08 |
| *** chandan_kumar has quit IRC | 20:10 | |
| clayg | ashish_: not sure - sorta doubt it, if you're doing loopback's at all it's sort of a weird hack and you have to watch out for weird subtle differences between dev and real physical multinode deploys - it depends a lot on what you're trying to elarn | 20:10 |
| clayg | ashish_: it *sounds* like maybe you want to see what a "real" capacity adjustment might be like - but in that case I'd recommend you pop a couple of vms with multiple virtual disks attached and configure a multi "node" swift environment, then add another vm or add some more virtualdisks to your existing vms. | 20:11 |
| ashish_ | clayg Cant I add disks on a single node setup? | 20:12 |
| *** csd has joined #openstack-swift | 20:12 | |
| clayg | ashish_: where a single node setup is a swift-all-in-one that pretends to be a mutinode setup (yes but it's very strange)? Or a real single node setup (yeah sure easy as pie)? | 20:13 |
| ashish_ | clayg you are right I wanted to see the capacity adjustments of swift | 20:13 |
| clayg | ashish_: k, well you should probably start with a multinode setup, the swift-all-in-one setup is not for that kind of work IMHO. | 20:13 |
| ashish_ | Okay clayg thanks. | 20:16 |
| clayg | you could sorta do a multi node setup (one set of storage servers per node) that just happens to have a single node, and then just add disks to that node - but I don't think that setup is strictly documented - it's just a really special case of a multi-node deploy | 20:17 |
| tdasilva | clayg, portante: how are tombstone files cleaned up? | 20:19 |
| clayg | tdasilva: object replciators | 20:20 |
| clayg | look for reclaim_age | 20:20 |
| ashish_ | clayg I would try the multinode and would contact you for further help.thanks . | 20:21 |
| tdasilva | clayg: got it, thanks! | 20:21 |
| clayg | ashish_: just ask in here, someone is normally around... if i'm online and I see it I'll throw in my two cents - i'm sort of a blow hard like that... | 20:22 |
| ashish_ | clayg: okay clayg thanks. | 20:25 |
| *** rustlebee is now known as russellb | 20:27 | |
| *** gyee has quit IRC | 20:57 | |
| *** mmcardle has joined #openstack-swift | 20:57 | |
| *** Trixboxer has quit IRC | 20:57 | |
| *** piyush2 has quit IRC | 21:01 | |
| *** mmcardle has quit IRC | 21:01 | |
| *** tdasilva has left #openstack-swift | 21:04 | |
| *** Guest__ has quit IRC | 21:24 | |
| *** Guest__ has joined #openstack-swift | 21:25 | |
| *** G________ has joined #openstack-swift | 21:30 | |
| *** Guest__ has quit IRC | 21:31 | |
| *** lpabon has quit IRC | 21:44 | |
| *** MooingLemur has joined #openstack-swift | 21:47 | |
| *** jasondotstar has joined #openstack-swift | 21:49 | |
| *** mmcardle has joined #openstack-swift | 21:52 | |
| *** ashish_ has quit IRC | 21:52 | |
| *** mmcardle has quit IRC | 21:53 | |
| *** zackf has quit IRC | 21:58 | |
| peluse | clayg: still there? | 22:05 |
| openstackgerrit | paul luse proposed a change to openstack/swift: Slight modification to the object updater sweep code https://review.openstack.org/87045 | 22:09 |
| clayg | peluse: regarding object updater change, i do see now that change is somewhat storage policy specific - i'll give it whirl after a bit tho | 22:11 |
| peluse | yeah, well, there it is :) | 22:11 |
| peluse | clayg: I can always retract it if we don't think its a good candidate right now.... | 22:12 |
| clayg | peluse: not sure - worth a shot ;) | 22:12 |
| *** gyee has joined #openstack-swift | 22:16 | |
| *** bobf has joined #openstack-swift | 22:16 | |
| *** mmcardle has joined #openstack-swift | 22:23 | |
| *** _bluev has joined #openstack-swift | 22:23 | |
| *** mmcardle has quit IRC | 22:28 | |
| zaitcev | clayg: I would be fine with test.conf being the ultimate too. Persuade Peter/. | 22:38 |
| clayg | zaitcev: he said he'd do it, poor guy got all ganged up on | 22:38 |
| *** bobf has quit IRC | 22:47 | |
| *** G________ has quit IRC | 22:47 | |
| *** mmcardle has joined #openstack-swift | 22:56 | |
| *** jasondotstar has quit IRC | 22:59 | |
| *** mmcardle has quit IRC | 23:02 | |
| *** _bluev has quit IRC | 23:15 | |
| *** dmsimard has joined #openstack-swift | 23:16 | |
| openstackgerrit | paul luse proposed a change to openstack/swift: Add Storage Policy Support to Container Sync https://review.openstack.org/86469 | 23:19 |
| openstackgerrit | Michael Barton proposed a change to openstack/swift: Sendfile acceleration for GET requests https://review.openstack.org/82186 | 23:26 |
| *** occupant has quit IRC | 23:37 | |
| *** piousbox has quit IRC | 23:47 | |
| *** mmcardle has joined #openstack-swift | 23:58 | |
| *** NM1 has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!