| *** NKleppan has quit IRC | 00:02 | |
| *** kei_yama has quit IRC | 00:02 | |
| *** kei_yama has joined #openstack-swift | 00:03 | |
| *** catintheroof has joined #openstack-swift | 00:11 | |
| *** klamath has quit IRC | 00:14 | |
| *** siva_krish has joined #openstack-swift | 00:49 | |
| *** gyee has quit IRC | 00:53 | |
| *** catintheroof has quit IRC | 00:54 | |
| *** chlong has quit IRC | 01:11 | |
| kota_ | good morning | 01:38 |
|---|---|---|
| *** zhugaoxiao has quit IRC | 01:47 | |
| *** siva_krish has quit IRC | 01:54 | |
| openstackgerrit | Merged openstack/swift: SLO: Concurrently HEAD segments https://review.openstack.org/391682 | 01:58 |
| mattoliverau | kota_: morning | 02:05 |
| *** takashi_ has joined #openstack-swift | 02:12 | |
| *** dmorita has quit IRC | 02:28 | |
| *** dmorita has joined #openstack-swift | 02:48 | |
| *** dmorita has quit IRC | 02:53 | |
| notmyname | kota_: thanks for the changelog review. I'm updating it now | 03:03 |
| kota_ | notmyname: sure, sorry I'm in later timezone than you :-) | 03:04 |
| notmyname | kota_: I'm the late one! you're in teh early timezone :-) | 03:04 |
| kota_ | hehe | 03:05 |
| *** tqtran has quit IRC | 03:07 | |
| notmyname | FWIW, if you have 2.11.0, save EC data, then roll back to 2.10.0, you cannot access the new data any more (404) because of the different filename | 03:09 |
| notmyname | but auditing doesn't remove it | 03:10 |
| notmyname | however, I'd expect that losing a drive while you're in a rollback state could degrade durability | 03:10 |
| notmyname | I'll make a note to not roll back. good catch, kota_ | 03:10 |
| kota_ | notmyname: thanks for making sure the behavior so quickly. | 03:11 |
| kota_ | that's while I was lunching | 03:12 |
| notmyname | kota_: so the pyeclib question is about landing the warning patch, tagging it, then bumping it in swift, then releasing swift? | 03:13 |
| kota_ | notmyname: correct, if you want. but i think it's ok we handle it independently | 03:13 |
| kota_ | notmyname: because, current requirements.txt is pyeclib>=1.3.1 so anytime we can find the log warning change when we bump the pyeclib version | 03:14 |
| kota_ | for clean deployment though. | 03:15 |
| notmyname | right | 03:15 |
| notmyname | I'd say let's do it, but that would mean also waiting on the global-requirements update before we can land a change to our repo's requirements.txt | 03:15 |
| kota_ | yeah, that's what I minded | 03:15 |
| kota_ | so i don't have strong opinion to jugde which 1.doing that or 2.release swift and bump pyeclib later for now. | 03:17 |
| notmyname | if we release swift now, then bump pyeclib, do you think we should do another swift release then, too? | 03:17 |
| notmyname | eg 2.11.1? | 03:17 |
| kota_ | hmm... i'm not sure we need warning log line so strongly | 03:19 |
| kota_ | the warning log line is a mitigation and temtative solution until packagers making liberasurecode >=1.3.1 | 03:19 |
| notmyname | really? I kinda feel like we should bump the requirements in swift asap and get that out to people | 03:19 |
| kota_ | hmm | 03:20 |
| notmyname | ok, if we wait to do a swift release, we're in the exact same situation as no swift release: people can still deploy a bad combination | 03:20 |
| kota_ | notmyname: that's too bad | 03:20 |
| notmyname | if we release swift now and later do a dep bump, then they get new good stuff now and can a bad combination | 03:20 |
| notmyname | I think releasing now with what we have, then doing a dependency bump is best | 03:21 |
| kota_ | ok, sounds great. | 03:21 |
| notmyname | and if we get it soon, then great. if not, that's too bad, but still better than just not releasing at all | 03:21 |
| kota_ | kk | 03:22 |
| kota_ | make sense | 03:23 |
| kota_ | notmyname: anything else i can help you for the release? | 03:23 |
| notmyname | thanks. I was just wondering about one thing related to the .durables change | 03:24 |
| notmyname | if we can't downgrade after writing data, how do rolling proxy upgrades work? there will be a situation where some proxies will have 2.10 and some will have 2.11. | 03:25 |
| notmyname | probably a good question for acoles_ | 03:25 |
| kota_ | i think the impl for non-.durable in the diskfile layer so the older proxy still can get conversation with the newer object servers which doesn't have .durable files. | 03:26 |
| notmyname | oh, right. it's all on the object layer. so that will be fine | 03:27 |
| kota_ | IIRC | 03:27 |
| notmyname | yeah, that makes sense. my mistake | 03:27 |
| kota_ | :-) | 03:29 |
| notmyname | kota_: I'm about to push a new version of the changelog. looking at it will help with the release. thanks :-) | 03:30 |
| notmyname | *about == after I get my kids in bed, so more like 30 min from now | 03:31 |
| kota_ | notmyname: ok. enjoy with your kids ;-) | 03:31 |
| *** sanchitmalhotra has joined #openstack-swift | 03:32 | |
| notmyname | no fussing, we're good :-) | 03:44 |
| notmyname | my kids do a lot of crazy things. but one thing they've always done well is bedtime | 03:44 |
| kota_ | :-) | 03:49 |
| openstackgerrit | John Dickinson proposed openstack/swift: 2.11.0 release notes https://review.openstack.org/399215 | 03:49 |
| notmyname | thre | 03:49 |
| notmyname | and this time I remembered the reno notes at the same time ;-) | 03:50 |
| kota_ | nice | 03:50 |
| *** dmorita has joined #openstack-swift | 03:50 | |
| *** klrmn has quit IRC | 03:54 | |
| *** dmorita has quit IRC | 03:55 | |
| *** psachin has joined #openstack-swift | 04:07 | |
| notmyname | kota_: thanks. I'll land it and prep the -release repo patch for permission to release | 04:23 |
| kota_ | notmyname: sounds good, thanks | 04:24 |
| *** takashi has joined #openstack-swift | 04:51 | |
| *** takashi_ has quit IRC | 04:52 | |
| *** dmorita has joined #openstack-swift | 04:52 | |
| *** dmorita has quit IRC | 04:57 | |
| *** ppai has joined #openstack-swift | 05:37 | |
| openstackgerrit | Merged openstack/swift: 2.11.0 release notes https://review.openstack.org/399215 | 05:56 |
| *** gabor_antal has quit IRC | 05:59 | |
| *** gabor_antal has joined #openstack-swift | 05:59 | |
| *** AbyssOne has quit IRC | 06:08 | |
| *** AbyssOne has joined #openstack-swift | 06:10 | |
| *** zaitcev has quit IRC | 06:16 | |
| *** mgagne has quit IRC | 06:18 | |
| *** sams-gleb has joined #openstack-swift | 06:18 | |
| *** timburke has quit IRC | 06:20 | |
| *** mgagne has joined #openstack-swift | 06:21 | |
| *** mgagne is now known as Guest52285 | 06:21 | |
| *** timburke has joined #openstack-swift | 06:23 | |
| *** ChanServ sets mode: +v timburke | 06:23 | |
| *** dmorita has joined #openstack-swift | 06:25 | |
| notmyname | FYI https://review.openstack.org/#/c/399387/ | 06:30 |
| patchbot | patch 399387 - releases - swift 2.11.0 release | 06:30 |
| *** dmorita has quit IRC | 06:30 | |
| *** sams-gleb has quit IRC | 06:32 | |
| *** jamielennox is now known as jamielennox|away | 06:36 | |
| *** sanchitmalhotra has quit IRC | 06:58 | |
| *** ChubYann has quit IRC | 07:04 | |
| *** tesseract has joined #openstack-swift | 07:18 | |
| *** tesseract is now known as Guest90313 | 07:18 | |
| *** klrmn has joined #openstack-swift | 07:26 | |
| openstackgerrit | howard lee proposed openstack/python-swiftclient: Replace 'assertEqual(None, ...)' with 'assertIsNone(...)' https://review.openstack.org/399410 | 07:28 |
| *** dmorita has joined #openstack-swift | 07:38 | |
| *** hseipp has joined #openstack-swift | 07:40 | |
| *** dmorita has quit IRC | 07:42 | |
| *** pcaruana has joined #openstack-swift | 07:45 | |
| *** jamielennox|away is now known as jamielennox | 08:03 | |
| *** openstackgerrit has quit IRC | 08:03 | |
| *** openstackgerrit has joined #openstack-swift | 08:03 | |
| *** amoralej|off is now known as amoralej | 08:12 | |
| *** klrmn has quit IRC | 08:15 | |
| *** hogepodge has quit IRC | 08:15 | |
| *** rledisez has joined #openstack-swift | 08:16 | |
| *** geaaru has joined #openstack-swift | 08:23 | |
| *** ayrus has joined #openstack-swift | 08:29 | |
| *** ayrus is now known as Guest32716 | 08:30 | |
| *** lifeless has quit IRC | 08:45 | |
| kota_ | sorry cschwede, i added to -2 to patch 399237 to hold the state. | 08:45 |
| patchbot | https://review.openstack.org/#/c/399237/ - swift - Avoid infinite loop while placing parts | 08:45 |
| kota_ | cschwede: probably we ware looking at the same patch at the same time. | 08:46 |
| *** lifeless has joined #openstack-swift | 08:47 | |
| cschwede | kota_: ok - what issue did you found? | 08:49 |
| kota_ | cscchwede: it may be small thing but i found the test doesn't fail w/o code change | 08:50 |
| *** dmorita has joined #openstack-swift | 08:50 | |
| kota_ | cschwede: not sure i'm addressing to the right thing though, not yet looking at deeply | 08:50 |
| kota_ | just starting to work | 08:50 |
| cschwede | kota_: the test runs infinitely w/o the code, i tested that - it just hangs | 08:51 |
| kota_ | cschwede: oh, really? that means, I'm missing something, my local testing results in all tests passed. | 08:52 |
| kota_ | w/o hangs | 08:52 |
| kota_ | test_round_off_error (test.unit.common.ring.test_builder.TestRingBuilder) ... ok | 08:53 |
| kota_ | hmmm | 08:53 |
| Guest32716 | hiii all, i am using my swift cluster from last one month and i performed several upload and delete of object operations | 08:54 |
| *** dmorita has quit IRC | 08:54 | |
| Guest32716 | but now i don't have any objects on my swift-cluster still the storage node is 23% filled | 08:55 |
| Guest32716 | is swift will occupy the storage after deleting of objects also? | 08:56 |
| kota_ | cschwede: ok. not yet reach out the reason why I didn't get failed but exactly the test for the rebalance looks to go infinite loop, https://gist.github.com/bloodeagle40234/e2ce7d55e140fdb88ff72a645cf7bcba | 09:01 |
| kota_ | interesting, via nosetests test/unit/common/ring/test_builder.py -> not stuck - via python test/unit/common/ring/test_builder.py -> stuck at the test | 09:09 |
| kota_ | caused from something like random seed??? | 09:09 |
| *** joeljwright has joined #openstack-swift | 09:20 | |
| *** ChanServ sets mode: +v joeljwright | 09:20 | |
| *** joeljwright1 has joined #openstack-swift | 09:22 | |
| kota_ | ugh!!! sorry cschwede, aboslutely my mistake found | 09:23 |
| cschwede | kota_: no worries, glad that you checked this! | 09:23 |
| *** joeljwright1 has quit IRC | 09:24 | |
| kota_ | cschwede: I attached to the test code but I had mistake I attached also the change... | 09:24 |
| *** joeljwright has quit IRC | 09:24 | |
| *** joeljwright has joined #openstack-swift | 09:25 | |
| *** ChanServ sets mode: +v joeljwright | 09:25 | |
| kota_ | and nosetests which refers repo inside catches the change but via python command it refers the installed dir outside of repo | 09:26 |
| kota_ | that is the reason why i got 2 different result from the same tests | 09:27 |
| kota_ | oops, gerrit seems to go odd state??? it gots +2 verified from jenkins but we don't have +A right now. | 09:28 |
| kota_ | can we kick jenkins via adding +A again? | 09:28 |
| cschwede | kota_: yes, seems so - jenkins is restarting | 09:31 |
| *** derekjhyang has quit IRC | 09:51 | |
| *** irclogger` has joined #openstack-swift | 09:54 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Don't overwrite built-in 'id' https://review.openstack.org/399484 | 09:54 |
| kota_ | cschwede: a follow up for patch 399237 | 09:55 |
| patchbot | https://review.openstack.org/#/c/399237/ - swift - Avoid infinite loop while placing parts | 09:55 |
| *** irclogger` has quit IRC | 09:55 | |
| *** sams-gleb has joined #openstack-swift | 10:00 | |
| *** acoles_ is now known as acoles | 10:06 | |
| *** ChanServ sets mode: +o acoles | 10:12 | |
| *** nikivi has joined #openstack-swift | 10:23 | |
| *** winggundamth has quit IRC | 10:28 | |
| Guest32716 | acoles are you there | 10:53 |
| *** dmorita has joined #openstack-swift | 10:54 | |
| acoles | Guest32716: hio | 10:57 |
| Guest32716 | have you seen my question which i posted at 14:25 | 10:58 |
| *** dmorita has quit IRC | 10:58 | |
| Guest32716 | acoles: have you seen my question which i posted at 14:25 | 11:00 |
| acoles | looking | 11:01 |
| acoles | 08:55 UTC ? | 11:02 |
| *** McMurlock1 has joined #openstack-swift | 11:02 | |
| acoles | Guest32716: swift stores "tomstone" files to record when an object was deleted so there will be some storage consumed even when no objects are in a cluster | 11:03 |
| Guest32716 | acoles: it was occupied 100GB in my storage node | 11:03 |
| acoles | these files should be cleaned up by the object auditor and/or replicator processes, but only after a reclaim_age period of time has passed, which IIRC is one week by default | 11:04 |
| acoles | also, if you had some drives offline when object were deleted then you may still have some replicas of objects on those drives. That's why the tombstone files are created on deletion, to supersede any stale replicas. | 11:06 |
| Guest32716 | acoles: thanks | 11:06 |
| *** m3m0_ has joined #openstack-swift | 11:06 | |
| Guest32716 | acoles: i have one more doubt regarding my uploading speed | 11:06 |
| *** McMurlock1 has quit IRC | 11:07 | |
| acoles | ...and, if the auditor ever found any corrupt data then it would move bad data file to quarantine directories on object servers which you'd need to clean up manually | 11:07 |
| Guest32716 | acoles: i am using swift cluster which has two nodes and those are connected with 10Gbps network and i am seeing 5.1GB file taking 35 secs | 11:08 |
| Guest32716 | is there any way to speed up my upload process | 11:08 |
| m3m0_ | hello everyone, does anyone know if it is normal for swift to have "occasionally" long delays in the responses? | 11:09 |
| Guest32716 | acoles: is there any parameter to 100% utilize all my cpu cores while uploading objects to swift? | 11:11 |
| *** nikivi has quit IRC | 11:22 | |
| *** kei_yama has quit IRC | 11:23 | |
| onovy | notmyname: hi, are you there? | 11:26 |
| *** caiobrentano_ has joined #openstack-swift | 11:31 | |
| Guest32716 | acloes: are you there? | 11:36 |
| *** caiobrentano_ has quit IRC | 11:38 | |
| *** caiobrentano_ has joined #openstack-swift | 11:39 | |
| openstackgerrit | Merged openstack/swift: Avoid infinite loop while placing parts https://review.openstack.org/399237 | 11:46 |
| *** dmorita has joined #openstack-swift | 11:50 | |
| openstackgerrit | Merged openstack/python-swiftclient: Replace 'assertEqual(None, ...)' with 'assertIsNone(...)' https://review.openstack.org/399410 | 11:53 |
| openstackgerrit | Merged openstack/swift: Don't overwrite built-in 'id' https://review.openstack.org/399484 | 11:53 |
| *** dmorita has quit IRC | 11:55 | |
| *** takashi has quit IRC | 12:03 | |
| acoles | Guest32716: could be that your 5GB put is disk io bound rather than CPU? | 12:03 |
| *** asettle has joined #openstack-swift | 12:17 | |
| *** asettle has quit IRC | 12:18 | |
| *** derekjhyang has joined #openstack-swift | 12:21 | |
| *** hogepodge has joined #openstack-swift | 12:31 | |
| acoles | cschwede: re patch 398631, afaict it fixes a regression that came since Mitaka - prior to Mitaka the response code was 412, then the bug caused 500, the patch takes us back to 412. I'm inclined to merge this and if there is a proposal to change 412 to 400 then consider it separately. Are you ok with that? | 12:34 |
| patchbot | https://review.openstack.org/#/c/398631/ - swift - Correctly send 412 Precondition Failed in copy mid... | 12:35 |
| cschwede | acoles: ah ok, yes, makes sense, let’s merge this | 12:37 |
| acoles | cschwede: ok, I'll add a not ein gerrit and +A | 12:37 |
| acoles | note* | 12:37 |
| cschwede | acoles: just did that :) | 12:37 |
| acoles | cschwede: thanks! | 12:38 |
| *** Guest32716 has quit IRC | 12:44 | |
| *** Jeffrey4l has quit IRC | 12:47 | |
| *** dmorita has joined #openstack-swift | 13:03 | |
| *** Jeffrey4l has joined #openstack-swift | 13:07 | |
| *** dmorita has quit IRC | 13:07 | |
| *** tqtran has joined #openstack-swift | 13:08 | |
| *** tqtran has quit IRC | 13:12 | |
| openstackgerrit | Merged openstack/swift: Correctly send 412 Precondition Failed in copy middleware https://review.openstack.org/398631 | 13:14 |
| *** ppai has quit IRC | 13:15 | |
| *** asettle has joined #openstack-swift | 13:24 | |
| *** silor has joined #openstack-swift | 13:25 | |
| *** asettle has quit IRC | 13:28 | |
| *** silor has quit IRC | 13:32 | |
| *** silor1 has joined #openstack-swift | 13:32 | |
| *** silor1 is now known as silor | 13:35 | |
| *** amoralej is now known as amoralej|lunch | 13:42 | |
| *** dmellado is now known as dmellado|lunch | 13:50 | |
| *** jaakkos has quit IRC | 14:10 | |
| *** dmellado|lunch is now known as dmellado | 14:12 | |
| *** daemontool has joined #openstack-swift | 14:13 | |
| *** siva_krish has joined #openstack-swift | 14:20 | |
| *** daemontool has quit IRC | 14:21 | |
| *** daemontool has joined #openstack-swift | 14:21 | |
| *** dmorita has joined #openstack-swift | 14:29 | |
| *** derekjhyang has quit IRC | 14:31 | |
| *** dmorita has quit IRC | 14:34 | |
| *** amoralej|lunch is now known as amoralej | 14:42 | |
| *** tqtran has joined #openstack-swift | 14:42 | |
| *** siva_krish has quit IRC | 14:43 | |
| *** chlong has joined #openstack-swift | 14:52 | |
| *** asettle has joined #openstack-swift | 14:54 | |
| *** klamath has joined #openstack-swift | 14:55 | |
| *** asettle has quit IRC | 14:58 | |
| notmyname | onovy: hello | 15:00 |
| *** klamath has quit IRC | 15:05 | |
| *** klamath has joined #openstack-swift | 15:05 | |
| *** sams-gleb has quit IRC | 15:06 | |
| notmyname | story on swift http://superuser.openstack.org/articles/globo-openstack-swift/ | 15:06 |
| *** sams-gleb has joined #openstack-swift | 15:06 | |
| tdasilva | lol, love the neymar stats, it has everything to do with the article | 15:08 |
| acoles | notmyname: is that the story that was in the magazine in Barcelona? | 15:08 |
| notmyname | yep | 15:08 |
| acoles | nice | 15:08 |
| notmyname | just published online today | 15:08 |
| acoles | notmyname: linkedin pointed me to a nice interview you had done on newton features | 15:09 |
| notmyname | the one by carol sliwa? | 15:09 |
| notmyname | techtarget? | 15:09 |
| acoles | yes. | 15:09 |
| notmyname | yeah, I liked how that turned out | 15:09 |
| *** klrmn has joined #openstack-swift | 15:10 | |
| acoles | the email subject line was like "John Dickinson in the news" and my first thought was "OMG what's happened" :D | 15:10 |
| notmyname | lol | 15:10 |
| *** sams-gleb has quit IRC | 15:11 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Make direct_client not override User-Agent https://review.openstack.org/399634 | 15:11 |
| *** klrmn has quit IRC | 15:11 | |
| tdasilva | hahaha | 15:11 |
| *** deanman has joined #openstack-swift | 15:14 | |
| deanman | Hello, I'm using swift CLI v3.0 to communicate with a OS Mitaka release but i get a authorization failure error. I've sourced the OS RC file and I'm able to use the openstack command. Is swift reusing these env. variables or do i have to pass them as arguements? | 15:16 |
| tdasilva | jrichli: good morning! Just saw your comment on patch 399275 and it was exactly what I was expecting. I think the challenging part with breaking down tests.py is trying to figure out what to do with TestCases that test two middlewares? | 15:18 |
| patchbot | https://review.openstack.org/#/c/399275/ - swift - breaking down tests.py into smaller pieces | 15:18 |
| tdasilva | e.g., in the case of TestSloWithVersioning should it go with test_versioned_writes or a future test_slo file? or yet maybe a third file? | 15:18 |
| tdasilva | I will also ask this same question on the patch, but wanted to bring it here first so that others could also weigh in | 15:19 |
| notmyname | so deanman had asked that question in #openstack, and I told him he should ask in here because I was about to step away. I'm hoping someone could help out | 15:20 |
| notmyname | FWIW, in the last few days swift's use of md5 in the ring and for etag has been raised as a potential blocker for deployers (to meet with checkbox compliance). seems that we're likely going to be asked about it again and to consider changing it | 15:22 |
| tdasilva | deanman: swift client will use certain env. variables if they are set, i believe they are OS_USERNAME, OS_TENANT_NAME, OS_AUTH_URL | 15:22 |
| *** siva_krish has joined #openstack-swift | 15:23 | |
| tdasilva | oh, and OS_PASSWORD | 15:23 |
| joeljwright | tdasilva: you beat me be 0.5s | 15:23 |
| joeljwright | s/be/by | 15:23 |
| tdasilva | deanman: oh, but if you really have more questions about the swift client joeljwright is the man to ask!!! ;) | 15:24 |
| joeljwright | deanman: happy to help if I can | 15:25 |
| acoles | joeljwright needs a faster internet connection ;) | 15:26 |
| tdasilva | hehe | 15:26 |
| *** sams-gleb has joined #openstack-swift | 15:26 | |
| deanman | Thank you, i have sousourced the RC file so I assume that they are used. Nevertheless i get this auth error. | 15:26 |
| deanman | I've even tried passing them as arguements, but i got the same response. | 15:27 |
| acoles | deanman: fwiw swiftclient also picks up OS_IDENTITY_API_VERSION | 15:27 |
| deanman | swift --auth $OS_AUTH_URL --os-username $OS_USERNAME --os-password $OS_PASSWORD stat --debug | 15:27 |
| deanman | acoles: That arguement is not exported in my RC file | 15:28 |
| deanman | is that something that an OS administrator should have setup up so that tenants that download RC files get them included? | 15:30 |
| acoles | assuming you are using keystone auth then you need --os-project-name (or have it in env) | 15:30 |
| joeljwright | deanman: are you using keystonev3? | 15:30 |
| deanman | joeljwright: I've sourced V2 RC file but i've also download V3. | 15:31 |
| acoles | and maybe joeljwright remembers but i think --os-identity-api-version=3 ? or 2.0 ? | 15:31 |
| deanman | Are you suggesting to try out v3 ? | 15:31 |
| joeljwright | deanman: the collection of auth env variables you need depends on the auth system you're using | 15:32 |
| joeljwright | I'm assuming you're using keystone auth | 15:33 |
| deanman | joeljwright: pardon my OS overall ignorance but in the OS i got access is called "Identity" | 15:34 |
| *** sgundur_ has joined #openstack-swift | 15:34 | |
| *** lcurtis has joined #openstack-swift | 15:35 | |
| acoles | clayg: anyone - can you think of any reason why this would break anything or be a bad idea - patch 399634 | 15:35 |
| patchbot | https://review.openstack.org/#/c/399634/ - swift - Make direct_client not override User-Agent | 15:35 |
| joeljwright | deanman: could you list the OS_* variables that are set in your environment and the error you receive when you try to use the swift client? | 15:36 |
| joeljwright | you need to have OS_AUTH_URL, OS_USERNAME, OS_PROJECT_NAME and OS_PASSWORD set (or passed as arguments) | 15:38 |
| *** asettle has joined #openstack-swift | 15:41 | |
| *** sgundur_ has quit IRC | 15:44 | |
| *** pcaruana has quit IRC | 15:50 | |
| jrichli | tdasilva: yes, the tests with combinations of features are hard to place. I have commented on the patch. | 15:52 |
| *** tqtran has quit IRC | 15:55 | |
| tdasilva | jrichli: ack | 15:55 |
| *** sgundur_ has joined #openstack-swift | 15:55 | |
| *** admin6_ has quit IRC | 15:55 | |
| *** sgundur_ has quit IRC | 15:59 | |
| deanman | joeljwright: http://paste.ubuntu.com/23495954/ | 16:01 |
| deanman | joeljwright: it seems like that the --os-identity-api-version made the difference and returned correctly | 16:01 |
| *** sams-gleb has quit IRC | 16:01 | |
| *** sams-gleb has joined #openstack-swift | 16:02 | |
| joeljwright | that's interesting, because if you have OS_PROJECT_NAME set it should default to v3 | 16:03 |
| joeljwright | I'm glad it's working for you now though | 16:03 |
| joeljwright | deanman: does `swift --os-identity-api-version 3` work? | 16:04 |
| deanman | joeljwright: http://paste.ubuntu.com/23495979/ | 16:05 |
| *** daemontool has quit IRC | 16:05 | |
| joeljwright | deanman: thanks for the feedback - it seems that the logic to choose the identity version isn't working quite right | 16:06 |
| deanman | joeljwright: http://paste.ubuntu.com/23495988/ | 16:06 |
| *** sams-gleb has quit IRC | 16:06 | |
| joeljwright | deanman: Ah! I see why | 16:07 |
| *** sams-gleb has joined #openstack-swift | 16:07 | |
| joeljwright | the default to v3 when we see PROJECT/DOMAIN was added in 3.1.0 | 16:08 |
| joeljwright | if you upgrade to 3.1.0 you should be able to just call `swift list` with those variables set in your env | 16:09 |
| *** Guest90313 has quit IRC | 16:09 | |
| acoles | joeljwright: ah! I was about to say shouldn't the presence of a domain arg force v3 | 16:11 |
| *** sams-gleb has quit IRC | 16:13 | |
| *** sams-gleb has joined #openstack-swift | 16:14 | |
| rledisez | notmyname: do you know if there is a time limit to register for the PTG? I can't find any informations about that on the website or eventbrite | 16:15 |
| deanman | joeljwright: well it seems that is not available on my distro (Ubuntu xenial) but i'll find how i can download latest. In the meantime what's the name of the equivalent OS_* var name for api i should set on my RC ? | 16:16 |
| *** sams-gleb has quit IRC | 16:17 | |
| *** sams-gle_ has joined #openstack-swift | 16:17 | |
| *** sams-gle_ has quit IRC | 16:19 | |
| *** sams-gleb has joined #openstack-swift | 16:19 | |
| joeljwright | deanman: I believe it's ST_AUTH_VERSION | 16:22 |
| joeljwright | let me have a quick check | 16:23 |
| notmyname | rledisez: no I don't know about any deadline for PTG registration. there might be a limit on tickets though. I don't know. | 16:23 |
| notmyname | rledisez: I will find out | 16:24 |
| *** sams-gleb has quit IRC | 16:24 | |
| rledisez | notmyname: AFAIK, yes, there is limit on number of tickets: 500 | 16:25 |
| notmyname | swift release done! https://review.openstack.org/#/c/399387/ | 16:25 |
| patchbot | patch 399387 - releases - swift 2.11.0 release (MERGED) | 16:25 |
| joeljwright | deanman: if you set ST_AUTH_VERSION=3 you should be able to run without setting the auth version as an option | 16:25 |
| bkeller` | it says there's 366 tickets left for the PTG | 16:26 |
| notmyname | rledisez: yeah, that's what I was just told. 500 cap, not close to that now, but will move to a waiting list mode if necessary | 16:27 |
| deanman | joeljwright: got it, cheers! | 16:29 |
| *** SkyRocknRoll has joined #openstack-swift | 16:29 | |
| rledisez | ok, thx for the info (i just saw where to find the number of remaining tickets :)) | 16:29 |
| notmyname | rledisez: from what I can tell, let me know if you'll be registering late, but even if you do, I should be able to make sure you have a ticket | 16:30 |
| rledisez | notmyname: ok, thx for that. i'll try to get approval for travel quickly | 16:31 |
| *** SkyRocknRoll has quit IRC | 16:32 | |
| notmyname | https://twitter.com/notmyname/status/799651922059792384 for those that tweet | 16:34 |
| *** sgundur_ has joined #openstack-swift | 16:34 | |
| *** psachin has quit IRC | 16:37 | |
| *** klrmn has joined #openstack-swift | 16:43 | |
| *** hseipp has quit IRC | 16:44 | |
| *** klrmn has quit IRC | 16:44 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:46 | |
| *** abalfour has quit IRC | 16:48 | |
| pdardeau | notmyname: regarding md5, the osic security team recently did a first-pass audit of a number of openstack projects, and use of md5 was their only ding on swift | 16:48 |
| *** abalfour has joined #openstack-swift | 16:49 | |
| hurricanerix | pdardeau are you talking about the etag? | 16:49 |
| *** hseipp has joined #openstack-swift | 16:50 | |
| pdardeau | hurricanerix: my impression was that it was just out of principle (anywhere in the project) | 16:50 |
| *** cbartz has left #openstack-swift | 16:51 | |
| *** tqtran has joined #openstack-swift | 16:52 | |
| hurricanerix | Well, as far as I know, md5 is much faster than better "cryptographically strong" hashing algorithms. So it depends on if what you are using it for. If the thing doesn't need to be cryptographically strong (because it is not a security feature), then md5 is fine. (and as far as I know, the etag is not meant to be a security feature). Having said that, I think maybe people were working on a way where there | 16:53 |
| hurricanerix | could be alternate hashes used. clayg might know more about that maybe though... | 16:53 |
| pdardeau | hurricanerix: went digging through emails to find the specifics "... identified a few possible issues like usage of md5 hashes. ... As a norm, the security team recommends against usage of insecure hash functions." | 16:54 |
| hurricanerix | I think etag is just meant to be a easy way for a client to tell if something changed. and it isn't even defined as "md5", just something that should be unique. (I think) | 16:55 |
| hurricanerix | We just happen to use a md5 for it. | 16:55 |
| pdardeau | i believe the mentioned security scan just did an automated scan through the source repos. | 16:56 |
| hurricanerix | But I would argue that a insecure hashing function is fine in the context of something that is not sensitive from a security standpoint. | 16:56 |
| hurricanerix | Ahh | 16:56 |
| *** tqtran has quit IRC | 16:56 | |
| pdardeau | or even crc in some cases | 16:57 |
| *** siva_krish has left #openstack-swift | 16:58 | |
| *** sgundur_ has quit IRC | 17:00 | |
| *** rickflare has quit IRC | 17:07 | |
| hurricanerix | pdardeau Yeah, I just re-read over the purpose of the Etag, and I don't see any reason using an md5 for it would be a problem. https://tools.ietf.org/html/rfc7232#section-2.3 | 17:07 |
| redbo | If they're smart, they're complaining about how we use md5 to pick the partition, not about etags. | 17:11 |
| *** klrmn has joined #openstack-swift | 17:12 | |
| *** rledisez has quit IRC | 17:12 | |
| redbo | Even though that's kind of pretend too. There should probably be a setting on the ring to use a more secure algorithm. | 17:12 |
| *** hseipp has quit IRC | 17:22 | |
| hurricanerix | redbo good point | 17:22 |
| *** sgundur_ has joined #openstack-swift | 17:25 | |
| *** oshritf has joined #openstack-swift | 17:27 | |
| *** klamath has quit IRC | 17:31 | |
| *** Guest52285 is now known as mgagne | 17:33 | |
| *** mgagne has quit IRC | 17:33 | |
| *** mgagne has joined #openstack-swift | 17:33 | |
| *** SkyRocknRoll has quit IRC | 17:46 | |
| *** dmorita has joined #openstack-swift | 17:48 | |
| *** dmorita has quit IRC | 17:50 | |
| *** dmorita has joined #openstack-swift | 17:50 | |
| *** tqtran has joined #openstack-swift | 17:53 | |
| *** tqtran has quit IRC | 17:58 | |
| *** SkyRocknRoll has joined #openstack-swift | 17:59 | |
| *** asettle has quit IRC | 18:06 | |
| notmyname | yeah, it's largely what pdardeau said | 18:11 |
| timburke | kota_: good call on 399484 -- thanks | 18:12 |
| notmyname | people have issues if `grep -r md5 /path/to/swift/source` returns anything | 18:12 |
| *** sams-gleb has joined #openstack-swift | 18:14 | |
| notmyname | IIRC torgomatic had a patch at some point where the ring algo what settable | 18:14 |
| notmyname | from my reading, BLAKE* hashing looks interesting. fast and secure | 18:14 |
| timburke | joeljwright: on the env-var version detection, you're thinking of the OS_*_DOMAIN_* options 00 if any of those are set, we'll automatically go v3 | 18:16 |
| timburke | just project isn't enough | 18:18 |
| timburke | notmyname: given https://review.openstack.org/#/c/395237/ as a model, i don't even think it would be very difficult to do... | 18:21 |
| patchbot | patch 395237 - swift - add byteorder information and logic to ring files | 18:21 |
| joeljwright | timburke: I didn't look at the change | 18:21 |
| onovy | notmyname: hi again :) i have question about stable branch | 18:22 |
| joeljwright | timburke: because the version was too old anyway | 18:22 |
| joeljwright | :( | 18:22 |
| notmyname | timburke: yeah, chanign the ring hash is likely fairly easy. changing the etag is likely close to impossible (without api version bumps) | 18:22 |
| notmyname | onovy: what's up? | 18:22 |
| onovy | you are responsible for it review of it, right? or every core, with standard process? | 18:22 |
| timburke | yup. etag's gonna be a pain to ever change | 18:23 |
| clayg | timburke: what about if we *just* did the `while parts > 0` change and the rounding in parts_wanted be what it may? I doubt it would effect the final placement? | 18:23 |
| notmyname | onovy: it's slightly different than standard. it's good to have reviews as normal, but I'm the only one on the swift team who can land them (stable maint team can too) | 18:23 |
| onovy | notmyname: because i cherry-picked two patches: https://review.openstack.org/#/c/399531/ + https://review.openstack.org/#/c/399533/ | 18:23 |
| patchbot | patch 399531 - swift (stable/newton) - Set owner of drive-audit recon cache to swift user | 18:23 |
| patchbot | patch 399533 - swift (stable/newton) - Fix signal handling for daemons with InternalClient | 18:23 |
| timburke | clayg: did ya try it? 'cause that's where i started, too, and then it failed verification | 18:23 |
| clayg | oh interesting - yeah I'll need to try it | 18:24 |
| onovy | newton version is somehow special for be, because it will land stretch stable. i want to have it bug free as much as possible | 18:24 |
| onovy | *for me | 18:24 |
| *** hseipp has joined #openstack-swift | 18:24 | |
| notmyname | onovy: cool. yeah, there's a few things proposed to stable/* that I hope to land today. (I had just cleared that queue yesterday!) | 18:25 |
| notmyname | onovy: and the other thing I'm wanting to do is actually tag something there | 18:25 |
| openstackgerrit | Merged openstack/swift: Respect server type for --md5 check in swift-recon https://review.openstack.org/240726 | 18:25 |
| onovy | notmyname: cool. is it possible to do release of stable branch then? i know it's not up to you, but stable maint. But i think you should say: yes, do it :) | 18:25 |
| onovy | ah, lol. | 18:25 |
| *** oshritf has quit IRC | 18:26 | |
| notmyname | onovy: yeah, AFAIK it's up to the stable maint team to release it. but also AFAIK, they've never actually done that for swift. so I figure we can do it | 18:26 |
| onovy | sounds good | 18:26 |
| clayg | timburke: floats fucking suck | 18:31 |
| timburke | clayg: yup | 18:32 |
| *** hseipp has quit IRC | 18:32 | |
| *** SkyRocknRoll has quit IRC | 18:36 | |
| *** joeljwright has quit IRC | 18:40 | |
| *** acoles is now known as acoles_ | 18:46 | |
| openstackgerrit | Ondřej Nový proposed openstack/swift: object-replicator cleanup https://review.openstack.org/391617 | 18:47 |
| openstackgerrit | Shashirekha Gundur proposed openstack/python-swiftclient: modify 'swift <sub_command> —help' display https://review.openstack.org/390119 | 18:58 |
| *** asettle has joined #openstack-swift | 19:08 | |
| *** geaaru has quit IRC | 19:13 | |
| *** sgundur_ has quit IRC | 19:19 | |
| openstackgerrit | Tim Burke proposed openstack/swift: Reduce backend requests for SLO If-Match / HEAD requests https://review.openstack.org/347538 | 19:27 |
| openstackgerrit | Tim Burke proposed openstack/swift: Confirm receipt of SLO PUT with etag https://review.openstack.org/390901 | 19:28 |
| *** tqtran has joined #openstack-swift | 19:29 | |
| *** zul has quit IRC | 19:30 | |
| notmyname | FYI https://review.openstack.org/#/c/394600/ has landed in infra, which means that approximately tomorrow we'll have an XFS TMPDIR for our tests | 19:34 |
| patchbot | patch 394600 - openstack-infra/project-config - enable an xfs tmpdir for swift tests (MERGED) | 19:34 |
| *** klamath has joined #openstack-swift | 19:36 | |
| openstackgerrit | Tim Burke proposed openstack/swift: SLO: Make etag and size_bytes fully optional https://review.openstack.org/391090 | 19:38 |
| *** klamath has quit IRC | 19:43 | |
| *** deanman has quit IRC | 19:43 | |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: Accept more types of input for headers/meta https://review.openstack.org/359477 | 19:47 |
| *** deanman has joined #openstack-swift | 19:48 | |
| *** zul has joined #openstack-swift | 19:49 | |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: Allow arbitrary content-types for directory markers https://review.openstack.org/260689 | 19:50 |
| *** amoralej is now known as amoralej|off | 19:51 | |
| *** silor has quit IRC | 19:53 | |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: On auth failure, ignore passed-in tokens and retry https://review.openstack.org/399756 | 20:01 |
| *** klrmn has quit IRC | 20:13 | |
| *** sgundur_ has joined #openstack-swift | 20:15 | |
| *** klamath has joined #openstack-swift | 20:19 | |
| *** klrmn has joined #openstack-swift | 20:21 | |
| *** sgundur_ has quit IRC | 20:24 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: breaking down tests.py into smaller pieces https://review.openstack.org/399275 | 20:24 |
| clayg | patches be flying! | 20:26 |
| timburke | clayg: the concurrent HEAD patch for SLO resulted in a couple merge conflicts for me, which caused me to go on a fix-all-the-conflicts spree | 20:27 |
| timburke | that last one's new, though! | 20:27 |
| timburke | er, not tdasilva's; that's a rebase. the auth failure one | 20:27 |
| *** sgundur_ has joined #openstack-swift | 20:28 | |
| timburke | or maybe not. i should give tdasilva more credit -- i bet he's addressing comments! | 20:29 |
| tdasilva | timburke: i was addressing jrichli comments, and now I just noticed that acoles_ also left comments - yay! | 20:30 |
| tdasilva | sorry for the noise | 20:31 |
| *** klamath has quit IRC | 20:34 | |
| *** asettle has quit IRC | 20:45 | |
| timburke | idk, i think if anyone should apologize, it's me. and i'm not gonna! | 20:47 |
| clayg | timburke: that's the spirit! | 20:47 |
| timburke | progress! | 20:47 |
| *** sgundur_ has quit IRC | 20:49 | |
| clayg | kota_: so to test isa_l_rs_cauchy you also have to add it in pyeclib yeah? what's the related change for that? | 20:50 |
| notmyname | clayg: https://review.openstack.org/#/c/393276/ | 20:50 |
| patchbot | patch 393276 - pyeclib - ISA-L Cauchy support | 20:50 |
| clayg | i just found it :) | 20:50 |
| notmyname | yeah, I had it right at the top of my screen when you asked | 20:51 |
| clayg | pretty annoying backend definitions have to be statically updated in two different projects :\ | 20:52 |
| notmyname | yes | 20:52 |
| notmyname | I've started thinking google's model of "everything in one repo" isn't too bad ;-) | 20:52 |
| *** chlong has quit IRC | 20:53 | |
| clayg | it's just that we're not independently invested in making a great erasure code *library* - we're just making a storage system | 20:54 |
| clayg | oh well | 20:54 |
| notmyname | yeah, I actually don't think we should include those in swift (other things though...) | 20:54 |
| notmyname | but combining pyeclib and liberasurecode might not be terrible | 20:55 |
| tdasilva | clayg, kota_ I remember on wed we were talking about keeping users from setting up systems with isa_l_rs_vand and m > 4 | 20:57 |
| tdasilva | i learned that ceph does the same, but they also limit k > 21 | 20:58 |
| tdasilva | https://github.com/ceph/ceph/blob/master/src/erasure-code/isa/ErasureCodeIsa.cc#L366,L380 | 20:58 |
| tdasilva | just FYI... | 20:58 |
| notmyname | interesting | 20:58 |
| notmyname | I remember tsg saying that you can't have total more than 32 | 20:58 |
| notmyname | but k less than 22 is new info | 20:58 |
| *** caiobrentano_ has quit IRC | 20:59 | |
| *** silor has joined #openstack-swift | 21:03 | |
| tdasilva | gotta run, have a good weekend everyone | 21:04 |
| jrichli | have a good weekend, tdasilva! | 21:04 |
| *** silor1 has joined #openstack-swift | 21:08 | |
| *** silor has quit IRC | 21:10 | |
| *** silor1 is now known as silor | 21:10 | |
| *** mweshi has joined #openstack-swift | 21:22 | |
| *** dmorita has quit IRC | 21:27 | |
| *** dmorita has joined #openstack-swift | 21:28 | |
| *** dmorita has quit IRC | 21:30 | |
| *** dmorita_ has joined #openstack-swift | 21:30 | |
| *** catintheroof has joined #openstack-swift | 21:31 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 21:33 |
| *** Jeffrey4l has quit IRC | 21:33 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 21:34 |
| *** sams-gleb has quit IRC | 21:42 | |
| *** sams-gleb has joined #openstack-swift | 21:42 | |
| *** sams-gleb has quit IRC | 21:43 | |
| *** sams-gleb has joined #openstack-swift | 21:43 | |
| *** sams-gleb has quit IRC | 21:44 | |
| *** siva_krish has joined #openstack-swift | 21:45 | |
| *** mweshi has quit IRC | 21:52 | |
| *** silor has quit IRC | 21:56 | |
| notmyname | timburke: FYI the errors in the gate like https://review.openstack.org/#/c/391090/ are my fault and will be fixed as soon as https://review.openstack.org/#/c/399765/ takes effect | 21:57 |
| patchbot | patch 391090 - swift - SLO: Make etag and size_bytes fully optional | 21:57 |
| patchbot | patch 399765 - openstack-infra/project-config - Fix syntax error for tox-with-tmpdir-passed-in (MERGED) | 21:57 |
| timburke | notmyname: yeah, i saw that the fix merged, thought that would be enough to resolve the gate. i can wait :-) | 21:58 |
| *** asettle has joined #openstack-swift | 22:03 | |
| *** mweshi has joined #openstack-swift | 22:04 | |
| *** mweshi has quit IRC | 22:07 | |
| *** vint_bra has quit IRC | 22:11 | |
| *** dmorita_ has quit IRC | 22:55 | |
| *** dmorita has joined #openstack-swift | 22:55 | |
| *** dmorita has quit IRC | 23:00 | |
| *** dmorita has joined #openstack-swift | 23:04 | |
| *** dmorita has quit IRC | 23:22 | |
| *** dmorita has joined #openstack-swift | 23:23 | |
| *** dmorita has quit IRC | 23:24 | |
| *** dmorita has joined #openstack-swift | 23:24 | |
| *** Jeffrey4l has joined #openstack-swift | 23:31 | |
| *** klrmn has quit IRC | 23:47 | |
| *** lcurtis has quit IRC | 23:50 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!