| portante | and about license headers | 00:00 |
|---|---|---|
| notmyname | portante: nope. creiht and I worked out the -info one, so jsut waiting for the zuul dance | 00:00 |
| portante | I get pep8 failures on my system with that | 00:00 |
| notmyname | hmm..I was hoping someone wouldn't say anything about that ;-) | 00:00 |
| portante | did I do that? | 00:00 |
| notmyname | :-) | 00:00 |
| * portante puts on his best erkle immitation | 00:01 | |
| notmyname | portante: in the -info patch? | 00:01 |
| portante | so you are not concerned about the pep8 issues? | 00:01 |
| portante | yes | 00:01 |
| portante | three failures | 00:01 |
| portante | and the fact that it only has 27% coverage | 00:02 |
| portante | ;) | 00:02 |
| * portante portante piles on | 00:02 | |
| *** RockKuo has joined #openstack-swift | 00:02 | |
| portante | just shoot me | 00:02 |
| portante | if you want, i can work on the coverage | 00:03 |
| notmyname | looks like the header thing needs to be fixed to pass jenkins | 00:03 |
| notmyname | portante: if you work on coverage can you also push the headers? | 00:05 |
| portante | yes | 00:05 |
| notmyname | portante: this is the last thing for icehouse (except my authors/changelog updates which I'm working on now) | 00:05 |
| notmyname | thanks | 00:05 |
| portante | welcome | 00:06 |
| notmyname | portante: I'll be around for about 1.5 hours, then out for the night | 00:06 |
| portante | k | 00:06 |
| *** mwstorer has quit IRC | 00:07 | |
| *** shri has quit IRC | 00:11 | |
| *** csd has quit IRC | 00:20 | |
| *** matsuhashi has joined #openstack-swift | 00:28 | |
| *** csd has joined #openstack-swift | 00:30 | |
| *** bill_az has quit IRC | 00:51 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Clean up swift-{account, container}-info https://review.openstack.org/84456 | 00:57 |
| notmyname | portante: looking | 00:58 |
| portante | not done yet | 00:58 |
| portante | almost | 00:58 |
| portante | just wanted to show the vector I am no | 00:58 |
| portante | on | 00:58 |
| portante | just writing the final set of unit tests | 00:58 |
| portante | got coverage for the rings now | 00:58 |
| *** csd has quit IRC | 00:59 | |
| notmyname | cool | 00:59 |
| notmyname | portante: reconcile the exception catching on the ring in info.py | 00:59 |
| notmyname | portante: line 47 and 165 | 01:00 |
| notmyname | portante: are you going to do any more unification on print_*_info()? | 01:00 |
| notmyname | portante: also, I think it needs copyright headers of some sort | 01:01 |
| *** dmsimard has quit IRC | 01:02 | |
| portante | notmyname: I was following this page: http://docs.openstack.org/developer/hacking/ | 01:02 |
| portante | is that not the up-to-date one? | 01:02 |
| notmyname | no idea | 01:02 |
| notmyname | I normally just copy/paste from another file ;-) | 01:03 |
| notmyname | portante: thanks for working on it | 01:04 |
| creiht | wait... I thought we moved the print code out so that it wouldn't be duplicated | 01:04 |
| portante | notmyname: regarding unification, I had thought about it, but was going to drive the coverage first | 01:04 |
| portante | creiht: coverage first, man! | 01:04 |
| portante | then de-dupe | 01:04 |
| creiht | heh | 01:04 |
| portante | at least that is what I am going to do | 01:04 |
| creiht | ok I'll quit looking then :) | 01:04 |
| portante | ;) | 01:05 |
| openstackgerrit | A change was merged to openstack/swift: Add "If-None-Match: *" support to PUT https://review.openstack.org/81646 | 01:05 |
| creiht | at least until you are done | 01:05 |
| notmyname | creiht: :-) | 01:05 |
| notmyname | creiht: while you're waiting for that.... | 01:05 |
| *** matsuhashi has quit IRC | 01:05 | |
| creiht | oh no | 01:05 |
| notmyname | creiht: can you give me more detail on the if-none-match use case you have? I'm working on the changelog right now | 01:05 |
| creiht | notmyname: I'm busy 3d printing stuff | 01:05 |
| creiht | :) | 01:06 |
| notmyname | cool | 01:06 |
| creiht | hehe | 01:06 |
| *** matsuhashi has joined #openstack-swift | 01:06 | |
| creiht | so this customer only uploads one version of an object | 01:06 |
| creiht | they absolutely only want to upload if it isn't already there | 01:06 |
| notmyname | ie name == unique object? | 01:06 |
| creiht | so they have been doing heads before put | 01:06 |
| creiht | yes | 01:06 |
| creiht | which performs horribly for them | 01:06 |
| creiht | of course | 01:06 |
| notmyname | as one would expect | 01:07 |
| creiht | so this is an attempt to get them the same functionality but faster | 01:07 |
| creiht | I was hoping to add etag support as well so that it would be robust | 01:07 |
| creiht | but then ran into too many hairy edge cases | 01:07 |
| notmyname | creiht: ya I can see how that's much harder. | 01:08 |
| notmyname | even so, it's interesting to think about because it's a new feature that has a very specialized use case | 01:08 |
| creiht | the downside is that for it to be effective, you have to use expect: 100-continue | 01:08 |
| *** keving has left #openstack-swift | 01:08 | |
| creiht | which requests doesn't support >:( | 01:08 |
| notmyname | it's not functionally different than what was available before, but we need to be careful not to imply that you can use it to make an atomic write to swift | 01:09 |
| notmyname | creiht: ya, that's a big negative to requests | 01:09 |
| creiht | absolutely | 01:09 |
| notmyname | augh!! | 01:09 |
| notmyname | *can't* *can't* use it for atomic writes | 01:09 |
| notmyname | see?! | 01:09 |
| creiht | It shoulf fulfill the same use case as doing a HEAD before PUT if you aren't comparing etags | 01:12 |
| openstackgerrit | A change was merged to openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/83493 | 01:12 |
| notmyname | creiht: ok. makes sense. that's what I expected you to say, but I wanted to check if there was anything else there | 01:12 |
| clayg | unless the original write went to handoffs and the primaryes are now back online - in that case the head would 404, 404, 404, 200 where as if-none-match will just 200/overwrite | 01:13 |
| notmyname | clayg: :-) | 01:14 |
| creiht | it is also assumed that the version they are PUTing is the version they want | 01:14 |
| creiht | so in this case, they would just be helping the cluster recover :) | 01:14 |
| clayg | ^ i like that wording!!! | 01:14 |
| notmyname | New feature: now you can help Swift recover from errors! | 01:15 |
| clayg | well.. not that wording the other part - swift's if-none-match support assumes the client is attempting to upload the version of data they want | 01:15 |
| creiht | notmyname: but yeah, it is a bit of an obscure feature :) | 01:16 |
| clayg | it's almost exactly the opposite of what you'd expect when sending if-none-match: * - but it's definately what we've got | 01:16 |
| notmyname | clayg: I don't agree with that | 01:16 |
| notmyname | or at least that phrasing | 01:17 |
| notmyname | I don't see it as the opposite of what I'd expect | 01:17 |
| * notmyname is hardly one to be determining what people not familiar with Swift are expecting | 01:17 | |
| clayg | i think if knew the data i had in hand was the version i wanted in the cluster i would say 'if-none-match: etag-of-data-about-to-be-uploaded' | 01:18 |
| clayg | if i wanted to upload some new version - but not overwrite some old version i don't know the content of - i'd say if-none-match: * | 01:18 |
| clayg | and that's just not really what you get *under failure* - the scary part is it works *most* of the time | 01:19 |
| clayg | i worry people will rely on the behavior because - "swift supports if-none-match" | 01:19 |
| creiht | clayg: depends on what if means | 01:19 |
| creiht | ;) | 01:19 |
| clayg | i just wish it was documented to be more explicit about what usecase it's appropriate for and the gaurantees we can acctually offer vs what will be observed most of the time but not guaranteed | 01:21 |
| *** nthacker has joined #openstack-swift | 01:22 | |
| nthacker | hi All! Im interested in trying out swift on filesystems w/o xattr support | 01:22 |
| nthacker | i have been digging through the code myself... | 01:22 |
| nthacker | and in diskfile.py i see read_metadata and write_metadata which seem to do this work (writing Metadata to a file) | 01:23 |
| nthacker | but when I try this on a filesystem over NFSv3 | 01:23 |
| nthacker | it doesnt seem to work... | 01:23 |
| clayg | YES! YES! YES! | 01:23 |
| nthacker | any tips? | 01:23 |
| * clayg goes to get popcorn | 01:23 | |
| nthacker | clayg: haha | 01:24 |
| creiht | I was worried you were going to try to run it on NTFS :) | 01:24 |
| swifterdarrell | nthacker: NFSv4 | 01:24 |
| nthacker | creiht: haha, no. i am trying to run it on ext4 over nfsv3 | 01:24 |
| nthacker | swifterdarrell: yes that works! | 01:25 |
| notmyname | swifterdarrell: dont' you mean NFSv4.1? | 01:25 |
| swifterdarrell | nthacker: also, you're better off with reiserfs over nfsv4 | 01:25 |
| nthacker | swifterdarrell: yes i realize nfsv4 is a good way to go, but lets say i really need to use nfsv3 :) ...im jus hacking around though | 01:25 |
| notmyname | nthacker: so here's the deal. it's really late in the day for all of us and you aren't getting serious answers. the serious answer is to not do that. | 01:25 |
| nthacker | swifterdarrell: i do see some work that will come out in icehouse | 01:26 |
| notmyname | seriously. don't do that (also, don't listen to swifterdarrell) | 01:26 |
| notmyname | what you want to do is wait | 01:26 |
| notmyname | like 6 months | 01:26 |
| nthacker | notmyname: wow... | 01:26 |
| notmyname | and we should have better interfaces in swift for that | 01:26 |
| notmyname | they're sorta there now | 01:26 |
| notmyname | but they'll get better | 01:26 |
| nthacker | notmyname: can i try some bleeding edge code?? | 01:27 |
| notmyname | and it will be a huge about of work to do it now | 01:27 |
| notmyname | which would be great and all, but really. it will be a. lot. of. work. | 01:27 |
| nthacker | notmyname: ah ....interesting....though the code looks quite simple at first glance, i'd take your word for it | 01:27 |
| swifterdarrell | nthacker: maybe embed the swift-proxy into a kernel module inserted just after nfsv3 | 01:27 |
| nthacker | swifterdarrell: haha | 01:28 |
| swifterdarrell | nthacker: just make sure to pay attention to If-None-Match headers, they may increase network filesystem latencies under load | 01:28 |
| swifterdarrell | also noatime | 01:28 |
| notmyname | nthacker: I really want to see swift supporting stuff like that. specifically, I really like the idea of migrating in existing data stores into swift. but the reality is that we arent' there yet | 01:28 |
| *** nosnos has joined #openstack-swift | 01:29 | |
| nthacker | notmyname: I understand...i read some of the blueprints and the refactorings for diskfile...but what if i wanted to hack something to make it *work* | 01:29 |
| notmyname | nthacker: and we've seen code and had conversations with existing storage vendors who want to make swift work with their existing products (you may work for one of those companies, I don't know) | 01:29 |
| notmyname | nthacker: then you're looking at the DiskFile abstraction | 01:29 |
| notmyname | nthacker: good places to start are with the gluster-swift project (which, despite its name, doesn't have much to do with glusterFS). it's about putting swift on top of an arbitrary file system | 01:30 |
| notmyname | nthacker: and as such it's a really interesting project. | 01:30 |
| nthacker | notmyname: Great! in gluster-swift they have managed to hack the DiskFile abstraction | 01:31 |
| nthacker | to work with fs w/o xattrs? | 01:31 |
| notmyname | not hack. use properly ;-) | 01:31 |
| notmyname | actually I'm not sure. gluster-swift today is designed to run swift code (upstream, no fork) on top of glusterfs | 01:31 |
| *** zackf has joined #openstack-swift | 01:32 | |
| nthacker | notmyname: well that sounds interesting - i'll def poke around their repo on github | 01:32 |
| notmyname | I've installed it. it takes about 10 minutes. they are providing their own disk file abstraction to make it work | 01:32 |
| swifterdarrell | nthacker: if you want a real project, make it store metadata in long cylinders filled with mercury | 01:32 |
| nthacker | swifterdarrell: some real innovative ideas you got! | 01:32 |
| nthacker | swifterdarrell: not trying to change the world you know, just store my metadata in a file :) | 01:33 |
| Alex_Gaynor | swifterdarrell: I'm totally working on a project whose tagline is "Mercury delay lines for the cloud" | 01:33 |
| notmyname | Alex_Gaynor: (he was looking for that phrase) | 01:33 |
| notmyname | swifterdarrell: what about memristors? | 01:33 |
| notmyname | memristors in the cloud! you can have docker in your storage system? | 01:34 |
| notmyname | portante: I'm going to need to check out soon. I'll look at the patch again in the morning, and we won't cut anything until it lands | 01:36 |
| notmyname | portante: but i'm happy with where it is and is going | 01:36 |
| portante | great, almost got the coverage | 01:36 |
| portante | notmyname: | 01:37 |
| notmyname | portante: cschwede: acoles: chmouel: (all you core people significantly east of me) it would be very nice to have it reviewed and merged as soon as possible tomorrow, since ttx is in france | 01:37 |
| notmyname | so that I can do the authors/changelog and we'll be ready asap | 01:38 |
| *** dmorita has joined #openstack-swift | 01:38 | |
| openstackgerrit | Pete Zaitcev proposed a change to openstack/swift: Pluggable Back-ends for account and container servers https://review.openstack.org/47713 | 01:41 |
| *** zackf has quit IRC | 01:44 | |
| *** tsg has quit IRC | 01:49 | |
| *** haomaiwang has quit IRC | 02:02 | |
| *** haomaiwang has joined #openstack-swift | 02:02 | |
| *** haomaiw__ has joined #openstack-swift | 02:20 | |
| *** zanc has quit IRC | 02:23 | |
| *** haomaiwang has quit IRC | 02:23 | |
| *** zaitcev has quit IRC | 02:39 | |
| *** bsdkurt has quit IRC | 02:53 | |
| *** bsdkurt has joined #openstack-swift | 02:54 | |
| *** fifieldt_ is now known as fifieldt | 02:54 | |
| *** madhuri has quit IRC | 02:58 | |
| *** matsuhashi has quit IRC | 03:05 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Clean up swift-{account, container}-info https://review.openstack.org/84456 | 03:08 |
| *** nosnos has quit IRC | 03:23 | |
| *** zackf has joined #openstack-swift | 03:39 | |
| *** Edward-Zhang has joined #openstack-swift | 03:53 | |
| *** zackf has quit IRC | 03:55 | |
| *** judd7_ has quit IRC | 03:55 | |
| *** madhuri has joined #openstack-swift | 04:04 | |
| *** Edward-Zhang has quit IRC | 04:07 | |
| *** nosnos has joined #openstack-swift | 04:09 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Move module level setup under setup_package method https://review.openstack.org/84339 | 04:09 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Only change collation for functional tests https://review.openstack.org/84325 | 04:10 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Rework use of constraints to ease testing https://review.openstack.org/84877 | 04:10 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: In-process swift server for functional tests https://review.openstack.org/66108 | 04:10 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Read the configuration once for all func tests https://review.openstack.org/84332 | 04:10 |
| hugokuo | gaob .... I think it's a bug in Swift | 04:14 |
| *** matsuhashi has joined #openstack-swift | 04:15 | |
| portante | clayg, notmyname, acoles, chmouel: the above series of patches moves us towards being able to run in-process functional tests | 04:15 |
| portante | however, the MAX_FILE_SIZE limit tests fail for some reason, which I have not been able to figure out yet | 04:16 |
| portante | the changes leading up to in-process hopefully make sense and consolidate the functional test environment a bit | 04:17 |
| *** Midnightmyth has joined #openstack-swift | 04:17 | |
| portante | glange, since you have looked at one of the early ones in the series, would you be willing to consider the others too? | 04:17 |
| *** psharma has joined #openstack-swift | 04:18 | |
| *** gvernik has joined #openstack-swift | 04:21 | |
| *** gvernik has left #openstack-swift | 04:23 | |
| hugokuo | gaob : https://bugs.launchpad.net/swift/+bug/1301728 | 04:35 |
| Anju | portante:, hugokuo : why content-length is showing always 0 in a head of any object ? | 04:36 |
| hugokuo | Anju: not my case tho : http://paste.openstack.org/show/74922/ | 04:39 |
| hugokuo | Content-Length: 88 | 04:39 |
| Anju | hugokuo: let me check | 04:40 |
| *** matsuhashi has quit IRC | 04:40 | |
| Anju | hugokuo: one thing more , if I am giving content-length =0 | 04:40 |
| Anju | in a put request of any object | 04:40 |
| Anju | Its going fine | 04:41 |
| Anju | but when I am giving any value other than 0 | 04:41 |
| Anju | then put request not completing successfully | 04:42 |
| Anju | hugokuo: http://paste.openstack.org/show/74925/ | 04:44 |
| *** matsuhashi has joined #openstack-swift | 04:46 | |
| Anju | hugokuo: please try with other user(non admin) | 04:50 |
| *** ppai has joined #openstack-swift | 04:52 | |
| hugokuo | Anju: if you give the content-length: 0 in PUT request, then the object will be an empty object in Swift. That's why you saw the length of HEAD in 0. That make sense for me tho. | 04:56 |
| hugokuo | You need to specify the correct content-length during PUT. | 04:56 |
| Anju | hugokuo: http://paste.openstack.org/show/74926/ | 04:57 |
| *** Edward-Zhang has joined #openstack-swift | 04:58 | |
| hugokuo | Anju: get it back. Check if that's an empty file.... | 04:58 |
| Anju | hugokuo: no | 04:59 |
| Anju | its not an empty file | 04:59 |
| hugokuo | Anju: I'm pretty sure it's an empty object hence the Etag is d41d8cd98f00b204e9800998ecf8427e in your head response tho. | 05:02 |
| Anju | hugokuo: http://paste.openstack.org/show/74927/ | 05:03 |
| Anju | see the diff | 05:03 |
| hugokuo | Anju: What's fun in this post ? It's different object from my point of view.... | 05:05 |
| Anju | hugokuo: first I put from admin account in a different container | 05:06 |
| Anju | the same file (.viminfo) | 05:06 |
| hugokuo | full curl command, please | 05:07 |
| Anju | hugokuo: yeah sorry | 05:09 |
| Anju | hugokuo: http://paste.openstack.org/show/74928/ | 05:10 |
| hugokuo | Anju: What's the commands that you upload these two objects ... | 05:11 |
| Anju | hugoko : I uploaded first swift -A http://127.0.0.1:8080/auth/v1.0 -U admin:admin -K admin upload anju .viminfo | 05:12 |
| Anju | for test2 swift -A http://127.0.0.1:8080/auth/v1.0 -U test2:tester2 -K testing2 upload anjua .viminfo | 05:13 |
| *** Midnightmyth has quit IRC | 05:18 | |
| *** cheri has joined #openstack-swift | 05:18 | |
| hugokuo | Anju: http://paste.openstack.org/show/74929/ everything works as expectation in my end. ...... | 05:22 |
| Anju | hugokuo: ohhh , then | 05:23 |
| Anju | something going wrong in my side | 05:23 |
| hugokuo | Anju: Well, I guess that you just upload the wrong file. Maybe you can check the log for more clues there. | 05:24 |
| Anju | hugokuo: one thing more , you said if I will provide correct size of a file | 05:24 |
| Anju | in the content length | 05:25 |
| Anju | in a put request | 05:25 |
| Anju | of any object then it will be fine ? right ? | 05:25 |
| *** bvandenh has joined #openstack-swift | 05:33 | |
| hugokuo | Anju: I use -T to specify the upload file. It works for whatever the content-lenght is... | 05:37 |
| Anju | are you using curl ? | 05:37 |
| Anju | hugokuo: it will be very helpful if you can paste | 05:38 |
| hugokuo | curl -i -X PUT http://localhost/v1/AUTH_test2/1/test1 -H "X-auth-token:AUTH_tkf13c2948382c43b0ae39bd5949c59570" -T 1 | 05:38 |
| *** haomaiw__ has quit IRC | 05:40 | |
| *** haomaiwang has joined #openstack-swift | 05:41 | |
| openstackgerrit | Yuan Zhou proposed a change to openstack/swift: Clean up swift-{account, container}-info https://review.openstack.org/84456 | 06:00 |
| *** haomai___ has joined #openstack-swift | 06:01 | |
| *** cheri has quit IRC | 06:03 | |
| *** haomaiwang has quit IRC | 06:05 | |
| *** psharma has quit IRC | 06:16 | |
| yuan | hi peter, is the copyright belongs to openstack foundation, sth like this Copyright (c) 2010-2012 OpenStack Foundation? | 06:20 |
| yuan | portante: the header for Intel is: Copyright 2014 Intel Corporation, All Rights Reserved. | 06:26 |
| openstackgerrit | Christian Schwede proposed a change to openstack/swift: Clean up swift-{account, container}-info https://review.openstack.org/84456 | 06:26 |
| cschwede | yuan: afaik you can use a header like this: http://paste.openstack.org/show/74931/ | 06:29 |
| cschwede | python26 test already passed :) now i hope we can merge this very soon :) | 06:30 |
| *** psharma has joined #openstack-swift | 06:31 | |
| yuan | cschwede: OK, I see it, so the new file will need to have a copyright header from the author's employer? | 06:32 |
| cschwede | yuan: it doesn't need it - but you are free to add one, depends on your employer. disclaimer: i am no lawyer ;-) | 06:33 |
| yuan | cschwede: ahh, I think it's OK to skip that :) | 06:35 |
| *** matsuhashi has quit IRC | 06:36 | |
| *** nshaikh has joined #openstack-swift | 06:37 | |
| cschwede | yuan: all right :) | 06:37 |
| cschwede | yuan: so python27, pep8 passed too, as soon as all tests passed i'll add my +2 to ithe patch | 06:37 |
| *** matsuhashi has joined #openstack-swift | 06:39 | |
| yuan | thanks a lot! it turns out to be a bigger refactoring than i plan to do initialy... | 06:44 |
| *** psharma has quit IRC | 07:03 | |
| *** zigo has quit IRC | 07:04 | |
| *** zigo has joined #openstack-swift | 07:08 | |
| *** psharma has joined #openstack-swift | 07:30 | |
| *** cheri has joined #openstack-swift | 07:46 | |
| *** mmcardle has joined #openstack-swift | 07:53 | |
| *** nacim has joined #openstack-swift | 07:55 | |
| *** mlipchuk has joined #openstack-swift | 08:03 | |
| *** tanee-away is now known as tanee | 08:12 | |
| *** tanee is now known as tanee-away | 08:12 | |
| *** tanee-away is now known as tanee | 08:13 | |
| *** tanee is now known as tanee-away | 08:13 | |
| *** tanee-away is now known as tanee | 08:13 | |
| *** fbo_away is now known as fbo | 08:16 | |
| openstackgerrit | Victor Stinner proposed a change to openstack/python-swiftclient: Fix test_multithreading on Python 3 https://review.openstack.org/84104 | 08:31 |
| *** foexle has joined #openstack-swift | 08:44 | |
| *** psharma has quit IRC | 08:57 | |
| *** matsuhashi has quit IRC | 09:07 | |
| *** matsuhashi has joined #openstack-swift | 09:12 | |
| *** madhuri has quit IRC | 09:15 | |
| *** Longgeek_ has joined #openstack-swift | 09:24 | |
| *** psharma has joined #openstack-swift | 09:46 | |
| *** Trixboxer has joined #openstack-swift | 09:51 | |
| *** mkollaro has joined #openstack-swift | 09:53 | |
| *** Longgeek_ has quit IRC | 09:56 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Clean up swift-{account, container}-info https://review.openstack.org/84456 | 09:58 |
| *** matsuhashi has quit IRC | 09:59 | |
| *** matsuhashi has joined #openstack-swift | 09:59 | |
| portante | cschwede, yuan: ^^^, thanks for the fix, the use of datatime.utcfromtimestamp() method is needed so that we get consistent results on systems that don't run under UTC (like my laptop) | 10:00 |
| portante | chmouel, acoles: ^^^ | 10:00 |
| *** fbo is now known as fbo_away | 10:09 | |
| *** mkollaro has quit IRC | 10:10 | |
| Anju | portante: hiii, when container_stat tables put_timestamp field value will change ? on every object upload ? | 10:17 |
| portante | Anju: good question, I'd have to look ... but in a bit | 10:19 |
| Anju | portante: will be thankful | 10:20 |
| *** dmorita has quit IRC | 10:22 | |
| *** chandankumar_ has joined #openstack-swift | 10:22 | |
| *** Longgeek_ has joined #openstack-swift | 10:43 | |
| *** Edward-Zhang has quit IRC | 10:46 | |
| *** dmsimard has joined #openstack-swift | 10:46 | |
| *** matsuhashi has quit IRC | 10:48 | |
| *** Longgeek_ has quit IRC | 10:50 | |
| *** matsuhashi has joined #openstack-swift | 10:55 | |
| *** tdasilva has joined #openstack-swift | 11:06 | |
| *** mlipchuk has quit IRC | 11:10 | |
| *** tanee is now known as tanee-away | 11:11 | |
| *** nacim has quit IRC | 11:14 | |
| *** matsuhashi has quit IRC | 11:15 | |
| *** matsuhashi has joined #openstack-swift | 11:15 | |
| *** matsuhashi has quit IRC | 11:19 | |
| *** ppai has quit IRC | 11:21 | |
| *** mkollaro has joined #openstack-swift | 11:21 | |
| *** matsuhashi has joined #openstack-swift | 11:21 | |
| *** zhiyan_ is now known as zhiyan | 11:22 | |
| *** jamie_h has joined #openstack-swift | 11:23 | |
| *** mlipchuk has joined #openstack-swift | 11:26 | |
| Anju | portante: ^^^ | 11:29 |
| *** ppai has joined #openstack-swift | 11:32 | |
| *** Manish_ has joined #openstack-swift | 11:33 | |
| portante | Anju: just getting through my morning routine here in EDT, so I'll be able to review and answer a bit later, thanks for your patience. | 11:35 |
| Anju | portante: thankyou , no problem :) | 11:36 |
| *** Manish__ has joined #openstack-swift | 11:38 | |
| *** Manish_ has quit IRC | 11:38 | |
| *** tanee-away is now known as tanee | 11:38 | |
| cschwede | portante: yeah you're right, utcfromtimestamp is the way to go. | 11:44 |
| cschwede | portante: according to https://github.com/openstack-dev/hacking/blob/master/HACKING.rst#openstack-licensing an Apache license only header is fine, so if Yuan and Madhuri are fine with that we could finish this | 11:44 |
| *** tanee is now known as tanee-away | 11:48 | |
| *** tanee-away is now known as tanee | 11:49 | |
| *** matsuhashi has quit IRC | 11:52 | |
| Manish__ | hugokuo: Hi.... DO you know that who and when updates the "PUT_timestamp" column of "container_stat" table? | 11:58 |
| *** tdasilva has quit IRC | 12:02 | |
| portante | cschwede: okay, I'll fix up the bin one's to have just the license, and we can add madhuri's in another patch | 12:03 |
| *** cheri has quit IRC | 12:04 | |
| *** tanee is now known as tanee-away | 12:04 | |
| *** tdasilva has joined #openstack-swift | 12:04 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Clean up swift-{account, container}-info https://review.openstack.org/84456 | 12:05 |
| portante | cschwede: ^^^ | 12:05 |
| *** mlipchuk has quit IRC | 12:05 | |
| cschwede | portante: thanks - I'll add my +2 to it, to me the patch looks good. Hopefully Jenkins agrees as well :) | 12:07 |
| portante | ;) | 12:07 |
| *** tdasilva has left #openstack-swift | 12:08 | |
| *** StevenK_ has joined #openstack-swift | 12:10 | |
| *** fbo_away is now known as fbo | 12:10 | |
| *** StevenK has quit IRC | 12:10 | |
| *** occupant has quit IRC | 12:10 | |
| *** MooingLemur has quit IRC | 12:11 | |
| *** cschwede has left #openstack-swift | 12:12 | |
| *** cschwede has joined #openstack-swift | 12:12 | |
| *** JuanManuelOlle has joined #openstack-swift | 12:13 | |
| *** marcusvrn has joined #openstack-swift | 12:13 | |
| *** cheri has joined #openstack-swift | 12:16 | |
| Manish__ | Can anyone help me here? | 12:19 |
| chandan_kumar | Manish__, please ask your question here. | 12:26 |
| Manish__ | chandan_kumar: Hi... | 12:27 |
| Manish__ | DO you know that who and when updates the "PUT_timestamp" column of "container_stat" table? | 12:27 |
| *** mlipchuk has joined #openstack-swift | 12:28 | |
| portante | Anju, Manish__: if clayg does not answer when he gets in, i'll be looking at that in a bit | 12:30 |
| *** Manish__ has quit IRC | 12:31 | |
| *** Manish_ has joined #openstack-swift | 12:33 | |
| Manish_ | chandankumar_: Sorry..i got disconnected | 12:33 |
| Manish_ | did you go my query? | 12:33 |
| cschwede | Manish_: there are several places that trigger an update; for example object requests, container requests, background daemons (replicators/auditors) | 12:42 |
| *** RockKuo has quit IRC | 12:43 | |
| cschwede | Manish_: that said all updates should be called from swift/container/backend.py | 12:43 |
| openstackgerrit | Donagh McCabe proposed a change to openstack/swift: Reclaim containers even if account db is reclaimed https://review.openstack.org/84696 | 12:44 |
| Manish_ | i could find the scenarios where object_count, byte_used, hash columns are being updated...but couldn't find any scenarion when put_timestamp column is getting updated | 12:45 |
| *** tanee-away is now known as tanee | 12:46 | |
| Manish_ | cschwede: ^^^^ | 12:47 |
| *** tanee is now known as tanee-away | 12:47 | |
| marcusvrn | portante: ping | 12:50 |
| *** nosnos has quit IRC | 12:50 | |
| cschwede | Manish_: so normally this is only set on creation of the table. but then it might be overwritten by a replication process | 12:50 |
| *** PradeepChandani has joined #openstack-swift | 12:51 | |
| *** tanee-away is now known as tanee | 12:51 | |
| cschwede | Manish_: that will be the case if the rowids of two DBs differ by more than 50%, and in that case the whole DB file will be synced using rsync. | 12:52 |
| *** tanee is now known as tanee-away | 12:52 | |
| cschwede | Manish_: hmm, it might also get synced by the replicator without using rsync | 12:53 |
| *** ppai has quit IRC | 12:53 | |
| *** tanee-away is now known as tanee | 12:54 | |
| *** psharma has quit IRC | 12:55 | |
| *** StevenK_ is now known as StevenK | 13:02 | |
| *** cheri has quit IRC | 13:05 | |
| marcusvrn | Can anyone help me? I was trying to disable the object_server replication but keeping the others two working with replication (account_server and container_server). Is it possible? | 13:07 |
| ctennis | sure, just stop the swift-object-replicator process | 13:11 |
| *** changbl has quit IRC | 13:11 | |
| *** mmcardle has quit IRC | 13:13 | |
| marcusvrn | but is possible to disable it in some config file? Or I need to kill swift-object-replicator process each time that I rejoin stack? | 13:18 |
| *** cheri has joined #openstack-swift | 13:18 | |
| *** chandankumar_ has quit IRC | 13:19 | |
| *** mmcardle has joined #openstack-swift | 13:40 | |
| cschwede | portante, acoles: just one more +2 needed: https://review.openstack.org/#/c/84456/ | 13:46 |
| gholt | marcusvrn: You can remove the [object-replicator] section from the object-server.conf config file. | 13:51 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Clean up swift-{account, container}-info https://review.openstack.org/84456 | 13:55 |
| portante | cschwede: sorry, I did not feel comfortable with adding the Intel licences since that code was only refactored into info.py and test_info.py | 13:55 |
| portante | best left without copyrights until madhuri can weigh in | 13:56 |
| *** pberis has quit IRC | 13:58 | |
| gholt | Hehe, I forgot that we at Rackspace should really edit all the files and plaster our copyright on most everything... joking, joking. You can't really "All Rights Reserved" either since it's explicitly unreserving rights with the Apache License. | 14:00 |
| *** pberis has joined #openstack-swift | 14:04 | |
| *** pberis has quit IRC | 14:04 | |
| portante | gholt: yeah, seriously, much of this code should be copyright Rackspace, because the foundation can't take copyrights | 14:04 |
| *** pberis has joined #openstack-swift | 14:05 | |
| *** mkollaro has quit IRC | 14:05 | |
| *** mkollaro1 has joined #openstack-swift | 14:05 | |
| *** mkollaro1 is now known as mkollaro | 14:05 | |
| *** occupant has joined #openstack-swift | 14:07 | |
| gholt | I don't think it was always that way. And honestly, I don't remember why they can't be granted the copyright either. It's actually fine if folks want to place their tags here and there. Credit where due. But it could make for some ugly source files over time. Hehe | 14:08 |
| *** mlipchuk has quit IRC | 14:11 | |
| cschwede | portante: ok, makes sense to me | 14:11 |
| *** cheri has quit IRC | 14:13 | |
| *** mlipchuk has joined #openstack-swift | 14:13 | |
| marcusvrn | gholt: nice!! thx | 14:16 |
| *** piyush has joined #openstack-swift | 14:20 | |
| *** zackf has joined #openstack-swift | 14:23 | |
| *** bill_az has joined #openstack-swift | 14:25 | |
| *** pberis has quit IRC | 14:26 | |
| *** chandan_kumar has quit IRC | 14:27 | |
| *** chandan_kumar has joined #openstack-swift | 14:27 | |
| marcusvrn | portante: I saw that the DiskFile and DatabaseBroker as a Public API blueprint had the priority change from medium to undefined...the diskfile API will be in the icehouse release? do you guys want to change it or sth? | 14:28 |
| *** tanee is now known as tanee-away | 14:29 | |
| *** tanee-away is now known as tanee | 14:29 | |
| creiht | yeah I wish they would just get rid of the copyrights at the tops of files | 14:29 |
| *** chandan_kumar has quit IRC | 14:30 | |
| *** tanee is now known as tanee-away | 14:32 | |
| *** tanee-away is now known as tanee | 14:36 | |
| *** foexle has quit IRC | 14:39 | |
| *** mkollaro1 has joined #openstack-swift | 14:39 | |
| *** mkollaro has quit IRC | 14:39 | |
| notmyname | good morning | 14:45 |
| creiht | http://saallergy.info/day/2014-04-03 | 14:46 |
| creiht | :( | 14:46 |
| acoles | cschwede: been away, will look at https://review.openstack.org/#/c/84456/ now | 14:46 |
| creiht | lol... that one patch has recieved more attention then any before :) | 14:47 |
| *** pberis has joined #openstack-swift | 14:47 | |
| *** Manish_ has quit IRC | 14:51 | |
| *** tongli has joined #openstack-swift | 15:15 | |
| *** tanee is now known as tanee-away | 15:17 | |
| *** tanee-away is now known as tanee | 15:19 | |
| *** tanee is now known as tanee-away | 15:20 | |
| *** tanee-away is now known as tanee | 15:20 | |
| *** Midnightmyth has joined #openstack-swift | 15:21 | |
| *** mlipchuk has quit IRC | 15:22 | |
| *** jamie_h has quit IRC | 15:22 | |
| creiht | portante: thanks for the added coverage for the -info patch | 15:24 |
| *** mlipchuk has joined #openstack-swift | 15:25 | |
| *** jamie_h has joined #openstack-swift | 15:25 | |
| *** zhiyan is now known as zhiyan_ | 15:25 | |
| *** MooingLemur has joined #openstack-swift | 15:29 | |
| *** mwstorer has joined #openstack-swift | 15:40 | |
| portante | creiht: welcome | 15:43 |
| notmyname | hear hear | 15:45 |
| creiht | notmyname: anything else that needs to be reviewed before making the RC? | 15:48 |
| notmyname | yup. just finishing up my changelog line for describing dfg's cors changes. then I'll submit that to gerrit and would appreciate another set of eyes on it :-) | 15:49 |
| portante | notmyname: sure | 15:50 |
| creiht | cool | 15:50 |
| portante | any folks want to put some eyes on the in-process functional test stuff? | 15:50 |
| creiht | portante: I'll look at that after lunch | 15:50 |
| portante | thanks | 15:50 |
| *** mlipchuk has quit IRC | 15:50 | |
| *** Anju5 has joined #openstack-swift | 15:52 | |
| cschwede | portante: you mean https://review.openstack.org/#/c/66108/ ? | 15:56 |
| *** changbl has joined #openstack-swift | 15:59 | |
| portante | cschwede: yes, and all its dependencies which make it easier to see the refactoring | 15:59 |
| cschwede | portante: ah yes, so that are 5 patches. ok, will start with the top patch later | 16:01 |
| portante | thanks | 16:01 |
| openstackgerrit | John Dickinson proposed a change to openstack/swift: updates for 1.13.1 release https://review.openstack.org/85088 | 16:04 |
| *** ifadams has joined #openstack-swift | 16:04 | |
| notmyname | creiht: portante: cschwede: can you take a look at that please? ^^ | 16:04 |
| *** csd has joined #openstack-swift | 16:05 | |
| cschwede | notmyname: found a typo in first line in https://review.openstack.org/#/c/85088/1/CHANGELOG … just saying ;-) | 16:05 |
| notmyname | bah! | 16:06 |
| openstackgerrit | John Dickinson proposed a change to openstack/swift: updates for 1.13.1 release https://review.openstack.org/85088 | 16:06 |
| notmyname | cschwede: there ;-) | 16:06 |
| * notmyname goes to check other speling things | 16:06 | |
| notmyname | ah. one more. (in my defense, it's from portante's commit message) | 16:07 |
| cschwede | notmyname: last line in https://review.openstack.org/#/c/85088/2/.mailmap - is that correct? | 16:07 |
| cschwede | notmyname: VirtualBox mail address? | 16:08 |
| notmyname | cschwede: actually, yes it is (assuming it's the same person) | 16:08 |
| openstackgerrit | John Dickinson proposed a change to openstack/swift: updates for 1.13.1 release https://review.openstack.org/85088 | 16:08 |
| *** Anju5 has quit IRC | 16:08 | |
| notmyname | cschwede: author email on https://review.openstack.org/#/c/78530/ | 16:09 |
| cschwede | notmyname: ah yes. well then… so i only had a very quick look, but i think patchset 3 looks good? sorry, I’m in a hurry, can have another look in two hours (if it is not merged then) | 16:11 |
| notmyname | cschwede: thanks | 16:11 |
| notmyname | acoles: done | 16:11 |
| portante | ship it | 16:11 |
| openstackgerrit | John Dickinson proposed a change to openstack/swift: updates for 1.13.1 release https://review.openstack.org/85088 | 16:11 |
| notmyname | portante: cschwede: patch set 4 has the s/and/an/ from acoles | 16:12 |
| *** jogo has left #openstack-swift | 16:12 | |
| notmyname | portante: cschwede: acoles: thanks | 16:13 |
| *** krtaylor has quit IRC | 16:13 | |
| portante | it'd be nice to get the info patch to land first | 16:13 |
| notmyname | 2nd inthe gate queue now. it should | 16:14 |
| portante | great | 16:14 |
| notmyname | both will be in 1.13.1. when they land, that will be the release | 16:14 |
| portante | awesome | 16:14 |
| notmyname | thanks everyone | 16:14 |
| portante | so then we can release 1.13.2 next week with the in-process functional test work. ;) | 16:15 |
| portante | ha! | 16:15 |
| notmyname | continual releases! | 16:15 |
| notmyname | (I actually kinda like that idea) | 16:15 |
| portante | agreed | 16:15 |
| *** csd has quit IRC | 16:20 | |
| *** csd has joined #openstack-swift | 16:21 | |
| *** csd has quit IRC | 16:22 | |
| *** fbo is now known as fbo_away | 16:26 | |
| *** mmcardle has quit IRC | 16:27 | |
| *** zaitcev has joined #openstack-swift | 16:29 | |
| *** ChanServ sets mode: +v zaitcev | 16:29 | |
| *** mmcardle has joined #openstack-swift | 16:29 | |
| *** tanee has quit IRC | 16:34 | |
| *** tanee has joined #openstack-swift | 16:36 | |
| *** nshaikh has quit IRC | 16:42 | |
| *** krtaylor has joined #openstack-swift | 16:50 | |
| *** mmcardle has quit IRC | 16:51 | |
| *** piyush has quit IRC | 16:57 | |
| acoles | portante: you around? | 17:05 |
| portante | yes | 17:05 |
| portante | acoles: | 17:05 |
| acoles | hi. quick clarification on https://review.openstack.org/#/c/84877/3 ... | 17:06 |
| portante | ah yes, the fun one | 17:06 |
| acoles | re. functional tests using load_constraint() vs importing from constraints, | 17:07 |
| portante | yes | 17:07 |
| portante | that is the tricky one | 17:07 |
| acoles | so, right now the tests seem to use value is in swift.conf rather than whats in test.conf, by virtue of reload_constraints() | 17:08 |
| acoles | and this changes to preferring test.conf values over swift.conf, correct? | 17:09 |
| portante | yes: so, first, this is trying to work in two different scenarios: | 17:09 |
| portante | 1. an SAIO where the tree containing the functional tests is co-located with the deployment itself | 17:09 |
| portante | 2. a tree containing the tests has no access to the /etc/swift directory of the installation (co-located or on another machine) | 17:10 |
| portante | so if a test.conf has a setting, we need to honor that above access to a /etc/swift/swift.conf because of scenario #2, as I understand it | 17:11 |
| portante | for scenario #1, we still honor it, but usually the /etc/swift/test.conf file in an SAIO does not have any constraints set because they can be picked up directly by the test environment | 17:11 |
| portante | so when the test/functional/* code was referencing constraints directly, a setup using different constraints might end up with bad results | 17:12 |
| portante | and that is actually still the case, as there are tests that assume certain sizes of objects will always be accepted | 17:12 |
| acoles | ok. understand the two scenarios. the subtle change i observe is in scenario 1. that before, IF i had values 'non-standard' values set in swift.conf, the func test would pick them up, regardless of what is in test.conf. Now, the test will use what is in test.conf? because load_constraint prefers test.conf? | 17:15 |
| portante | only if the test.conf had that constraint set, yes | 17:16 |
| acoles | ok. i'm not saying that is wrong. i just wanted to make sure i understood the subtle shift in precedence. | 17:17 |
| portante | yes, now the behavior is consistent across the board, because some constraints before this change were using load_constraint and others were not | 17:18 |
| acoles | yeah. that is confusing :) | 17:18 |
| portante | so all this is needed for the in-process tests so that constraint references are not bound at module import time, or referenced from the wrong environment | 17:19 |
| acoles | yes. i see the dependency chain :) | 17:20 |
| portante | ;) | 17:21 |
| acoles | portante: i'll try to get review done before i leave. sorry i cannot contribute more to the cause today. | 17:21 |
| portante | that is okay, thanks for the help | 17:21 |
| *** tanee is now known as tanee-away | 17:22 | |
| *** judd7 has joined #openstack-swift | 17:23 | |
| torgomatic | so call me stupid (which is probably true right now) but why not use the /info response for the functest constraints? | 17:29 |
| torgomatic | that way you don't have to match what the proxy process does; you just ask it what it knows | 17:30 |
| *** tdasilva has joined #openstack-swift | 17:38 | |
| portante | ah, yes, we should do that | 17:42 |
| portante | err, I can make that happen | 17:42 |
| portante | torgomatic ... stayin' sharp ! | 17:43 |
| torgomatic | that's probably all the utility you'll get out of me today :) | 17:45 |
| *** mmcardle has joined #openstack-swift | 17:51 | |
| acoles | torgomatic: portante: nice idea. there's a couple of constraints i don't see in info e.g. max_header_size | 17:51 |
| *** mmcardle has quit IRC | 17:56 | |
| portante | torgomatic: all the best to you | 17:58 |
| portante | acoles: I'll take a look at reconciling that | 17:58 |
| *** mmcardle has joined #openstack-swift | 17:59 | |
| torgomatic | portante: thanks :) | 18:01 |
| *** Longgeek_ has joined #openstack-swift | 18:07 | |
| *** Longgeek_ has quit IRC | 18:12 | |
| *** ifadams has left #openstack-swift | 18:20 | |
| notmyname | snafus getting to work, but I'm back online now | 18:21 |
| *** mmcardle has quit IRC | 18:23 | |
| *** pberis has quit IRC | 18:27 | |
| *** pberis has joined #openstack-swift | 18:28 | |
| creiht | torgomatic: don't you have other things you need to be attending to? :) | 18:28 |
| *** mwstorer has quit IRC | 18:29 | |
| notmyname | so we're changing func tests to use /info to get constraints? | 18:29 |
| portante | creiht: it is hard switching from baby sitting one group to another group, let it go, let it go, ... | 18:29 |
| portante | notmyname: would like to | 18:29 |
| notmyname | that sounds like a great idea | 18:30 |
| portante | not in this patch series, since it will make it more complex | 18:30 |
| notmyname | and falling back to swift.conf defaults for constraints if /info doesn't respond? | 18:30 |
| portante | or test.conf | 18:32 |
| *** lpabon has joined #openstack-swift | 18:33 | |
| notmyname | right. test.conf defaults just mirror swift.conf defaults though | 18:36 |
| notmyname | test.conf exists because functests can do remote testing and there used to be no way to discover the constraints on the system | 18:36 |
| notmyname | hysterical raisins, and all that | 18:36 |
| *** jamie_h has quit IRC | 18:52 | |
| cschwede | portante: so i started reviewing https://review.openstack.org/#/c/84877/. what do you think about adding other default values as well to constraints? | 19:00 |
| portante | sec, will be available in a minute | 19:01 |
| cschwede | portante: no hurry | 19:01 |
| *** Midnightmyth has quit IRC | 19:03 | |
| notmyname | cschwede: I think some of the other constraints would be good, but probably should be added in their respective namespaces. eg https://github.com/openstack/swift/blob/master/swift/common/middleware/slo.py#L788 | 19:06 |
| notmyname | also, it looks like dynamic large objects aren't in /info | 19:07 |
| cschwede | notmyname: sorry, no i meant swift.common.constraints -> https://review.openstack.org/#/c/84877/3 | 19:08 |
| cschwede | notmyname: but i agree we should add missing constraints to /info | 19:08 |
| notmyname | cschwede: yup, actually that's what I was responding to. at least for stuff like max_large_object_get_time | 19:09 |
| portante | cschwede: would you like to propose a patch? | 19:09 |
| notmyname | cschwede: also, IMO /info should only expose stuff that affects the client API. ie not auditing parameters | 19:10 |
| cschwede | portante: for https://review.openstack.org/#/c/84877/3 or the /info thing? | 19:10 |
| notmyname | cschwede: unless those are in the /info "admin" request. then that's ok | 19:10 |
| cschwede | notmyname: yep, makes sense | 19:10 |
| notmyname | cschwede: the question you left on line 47 in that patch | 19:10 |
| cschwede | portante: notmyname: ok, i’ll propose a patch. should i put it in the existing from portante or a new one? | 19:11 |
| portante | do a new one, if you could | 19:11 |
| portante | that way it can move forward unemcombered by this work | 19:11 |
| cschwede | portante: all right, let’s do it like this. I will start with the patch now, but upload it tomorrow morning - i want to avoid silly patchsets from my side because of my local time ;-) | 19:12 |
| notmyname | enemcombered? unem*brush*ed? | 19:12 |
| *** lpabon has quit IRC | 19:14 | |
| zaitcev | In the traditional functests, is there a quick fix to make https work? Like an environment variable? | 19:19 |
| notmyname | zaitcev: it would be whatever returned as the storage url for the creds given, right? | 19:20 |
| zaitcev | notmyname: yes, mine returns https | 19:20 |
| notmyname | zaitcev: what's breaking? | 19:20 |
| zaitcev | (cd test/functionalnosetests && nosetests test_container.py:TestContainer.test_cross_account_container) | 19:21 |
| zaitcev | sec, I'll paste for clarity | 19:21 |
| zaitcev | http://paste.openstack.org/show/75014/ | 19:22 |
| zaitcev | this used to work, but of course the client is new now, it needs "insecure" passed to new connections | 19:22 |
| zaitcev | waaait a moment | 19:23 |
| zaitcev | I have a .pyc but no .py | 19:23 |
| zaitcev | the whole test/functionalnosetests/ was removed | 19:24 |
| notmyname | yes | 19:24 |
| notmyname | well, moved | 19:24 |
| *** mmcardle has joined #openstack-swift | 19:24 | |
| openstackgerrit | A change was merged to openstack/swift: updates for 1.13.1 release https://review.openstack.org/85088 | 19:25 |
| *** finite has joined #openstack-swift | 19:26 | |
| *** a_hacker_sboss has joined #openstack-swift | 19:27 | |
| *** finite has quit IRC | 19:27 | |
| *** a_hacker has joined #openstack-swift | 19:28 | |
| *** mmcardle has quit IRC | 19:28 | |
| *** fbo_away is now known as fbo | 19:29 | |
| zaitcev | anyhow, http://paste.openstack.org/show/75017/ | 19:30 |
| zaitcev | the problem is parsed[use_account], conn[use_account] = http_connection(url[use_account]) | 19:30 |
| zaitcev | for this server to work at all, http_connection() needs some insecure=True | 19:31 |
| notmyname | ya. probably will need some patches (rather than just an env var) | 19:32 |
| creiht | ugh | 19:33 |
| creiht | I'm really starting to dislike requests | 19:33 |
| *** piyush has joined #openstack-swift | 19:33 | |
| zaitcev | so... how do I set that, in a test? What's the preferred way? I could just hardcode it in test/functional/swift_testing.py, but that seems dumb. | 19:33 |
| creiht | http://docs.python-requests.org/en/latest/user/advanced/#blocking-or-non-blocking | 19:33 |
| creiht | swift-bench uses python-swiftclient, and it looks like concurrent perf has really degraded | 19:34 |
| notmyname | creiht: ugh | 19:34 |
| creiht | since swift-bench is using eventlet | 19:34 |
| creiht | still doing some testing, but I'm pretty sure that is the case | 19:35 |
| notmyname | creiht: FYI, short term solution maybe? https://github.com/swiftstack/ssbench/blob/master/requirements.txt | 19:35 |
| creiht | lol | 19:36 |
| creiht | well there is https://pypi.python.org/pypi/erequests | 19:36 |
| creiht | though it doesn't support the same api :/ | 19:37 |
| *** a_hacker2 has joined #openstack-swift | 19:37 | |
| notmyname | creiht: I'm not suggesting that ssbench replace swift-bench, just a short term solution for whatever testing you're doing | 19:37 |
| creiht | ahh | 19:38 |
| creiht | nah we are good | 19:38 |
| notmyname | so you can actually get stuff done today | 19:38 |
| notmyname | ah, ok | 19:38 |
| creiht | it gave pandemicsyn an excuse to play with golang more :) | 19:38 |
| notmyname | ie it's easier to write an ssbench scenario than it is to fix requests ;-) | 19:38 |
| *** wasmum has joined #openstack-swift | 19:38 | |
| notmyname | lol | 19:38 |
| creiht | just going back now and trying to figre out why swift-bench is broken | 19:38 |
| a_hacker2 | notmyname: greetings, I'm trying to continue some troubleshooting you did with some of my collegues last night. Do you have minute to look at a pastebin? I have an example of a container that just up and "disappears" | 19:38 |
| a_hacker_sboss | we were troubleshooting the issue as a 404, but have now determined we are dealing with "disappearance" | 19:39 |
| a_hacker_sboss | swift is hard ;) | 19:40 |
| a_hacker | So hard | 19:40 |
| creiht | how many a_hackers do we have on this ship? | 19:40 |
| creiht | :) | 19:40 |
| a_hacker2 | YO! | 19:40 |
| notmyname | he's multiplying! | 19:40 |
| a_hacker_sboss | the great ship titanic ;)) | 19:41 |
| a_hacker2 | I knew it I'm surrounded by hackers | 19:41 |
| *** Chet_McHacker has joined #openstack-swift | 19:41 | |
| notmyname | a_hacker(*): so what's up? | 19:41 |
| a_hacker2 | http://paste.openstack.org/show/y4iuUcytY9FFlEwfeUR2/ | 19:41 |
| a_hacker_sboss | lol ;) | 19:41 |
| a_hacker2 | appreciate your help notmyname | 19:42 |
| portante | any other takers for reviewing "Rework use of constraints to ease testing"? https://review.openstack.org/84877 | 19:43 |
| portante | Got one +2 so far ... ;) | 19:43 |
| cschwede | portante: you want to have an approve as well, right? ;-) | 19:44 |
| portante | yes! | 19:44 |
| glange | portante: I looked at that but I thought you were going to take the json stuff out? | 19:46 |
| portante | well, acoles just approved it, so if others don't mind taking that in, i'll leave it as is | 19:46 |
| portante | if other folks want that out, I'll take that out | 19:46 |
| portante | kinda like fishing | 19:46 |
| glange | oh, ok | 19:46 |
| glange | :) | 19:47 |
| portante | I got a fish on the hook | 19:47 |
| portante | just not in the boat | 19:47 |
| portante | so you can say that I am angling | 19:47 |
| portante | ;) | 19:47 |
| notmyname | a_hacker: how many account servers do you have running? | 19:47 |
| glange | portante: it looked good to me, but I probably agree with Zaitcev | 19:48 |
| notmyname | a_hacker: looks like just one? | 19:48 |
| *** a_hacker_sboss has quit IRC | 19:48 | |
| portante | glange: okay, i'll pull that out | 19:48 |
| portante | thanks | 19:48 |
| cschwede | portante: hmm, i think your patch is fine | 19:48 |
| portante | fight! | 19:48 |
| glange | merge it! :) | 19:48 |
| portante | cschwede: right now, it is 2 to 1 | 19:49 |
| cschwede | got me! | 19:49 |
| portante | ;) | 19:49 |
| wasmum | notmyname account servers = 1 | 19:49 |
| a_hacker2 | notmyname: 2 | 19:49 |
| a_hacker2 | 2 workers | 19:49 |
| Chet_McHacker | 2 workers | 19:49 |
| cschwede | another hacker! | 19:50 |
| notmyname | 2 more | 19:50 |
| Chet_McHacker | :O | 19:50 |
| *** wasmum has left #openstack-swift | 19:50 | |
| portante | so that means patch 84325 can be approved, too! | 19:50 |
| *** zhiyan_ is now known as zhiyan | 19:50 | |
| * portante wonders where this patch bot went to ... | 19:50 | |
| portante | patchbot: patch 84325 | 19:50 |
| patchbot | portante: https://review.openstack.org/#/c/84325/ | 19:50 |
| zaitcev | My logic was that as long as utils pretend to be like normal json then it should continue looking that way. If we had json.our_special_and_incompatible_json_loads(), then slapping utils on it would do no harm. | 19:50 |
| portante | ah there | 19:50 |
| cschwede | portante: yep, just looking at it | 19:51 |
| *** a_hackers_boss has joined #openstack-swift | 19:51 | |
| *** not_a_hacker has joined #openstack-swift | 19:52 | |
| *** not_a_hacker has left #openstack-swift | 19:52 | |
| *** a_not_hacker has joined #openstack-swift | 19:53 | |
| cschwede | since https://review.openstack.org/#/c/84332/ is already approved i will continue with https://review.openstack.org/#/c/84332/ tomorrow morning | 19:53 |
| notmyname | how many people do we have here for this? who's actually working on it that I shoudl be talking to? | 19:53 |
| notmyname | no need for all of Seagate to be here on it (dig -x says a seagate IP block) | 19:53 |
| portante | notmyname: any thing I can jump in on and help? | 19:54 |
| portante | maybe match the number of seagate engineers? ;) | 19:54 |
| notmyname | :-) | 19:54 |
| openstackgerrit | A change was merged to openstack/swift: Clean up swift-{account, container}-info https://review.openstack.org/84456 | 19:54 |
| portante | yeah! | 19:54 |
| portante | 1.13.1 is in the can | 19:55 |
| cschwede | i hear 1.13.1 coming :) | 19:55 |
| notmyname | a_hacker: can you add an "X-Newest: True" header to the listing request? I suspect it's likely that all replicas aren't getting updated | 19:55 |
| notmyname | portante: yay | 19:55 |
| a_hacker2 | ypu | 19:56 |
| a_hacker2 | yup | 19:56 |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Handle getting info on wrong database type https://review.openstack.org/85156 | 19:56 |
| *** mkollaro1 has quit IRC | 19:56 | |
| notmyname | a_hacker2: and since the proxy (by default) chooses a replica at random, you are getting inconsistent results | 19:57 |
| notmyname | or eventaully consistent results ;-) | 19:57 |
| portante | notmyname: above info patch, 85156 is a nicety, where if you user swift-container-info on an account .db file, by mistake, it won't give you a stack trace, but a simple message suggesting you might want to use another command | 19:58 |
| portante | in so many words, of course | 19:58 |
| notmyname | patchbot: p 85156 | 19:58 |
| patchbot | notmyname: https://review.openstack.org/#/c/85156/ | 19:58 |
| portante | just did not want to hold up 1.13.1 to wait for it | 19:58 |
| cschwede | portante: ah nice, was thinking the same this morning. | 19:58 |
| portante | I actually did that testing the info patch! | 19:58 |
| notmyname | ah, nice | 19:59 |
| portante | that will make our customers want 1.13.2 all the more. :) | 19:59 |
| portante | 1.13.1 is already passe | 19:59 |
| portante | ;) | 20:00 |
| portante | trunk chasers unite! | 20:00 |
| cschwede | you know how to keep the demand for new releases up ;) | 20:00 |
| portante | :) | 20:01 |
| *** mkollaro has joined #openstack-swift | 20:01 | |
| notmyname | now to finish the traditional post-facto launchpad updates... | 20:01 |
| portante | kewl | 20:02 |
| notmyname | ok, https://launchpad.net/swift/+milestone/1.13.1-rc1 shouldn't be as horribly out of date as it was | 20:02 |
| a_hacker2 | notmyname: http://paste.openstack.org/show/fqzfD5YyzDWWNXookJce/ | 20:02 |
| notmyname | a_hacker2: I'm trying to work though some edge cases in my mind. I'm not used to thinking about what happens with only one server in the system | 20:04 |
| a_hacker2 | we went down to one server for troubleshooting - it has been happening on a 8 storage node cluster | 20:04 |
| notmyname | a_hacker2: should still be fine since there are 3 replicas and more than 3 drives | 20:05 |
| notmyname | a_hacker2: what version of swift are you using? | 20:05 |
| a_hacker2 | it's a havana release: >>> swift.__canonical_version__ '1.10.0' | 20:06 |
| notmyname | ok | 20:06 |
| notmyname | a_hacker2: at this point, I'd use swift-get-nodes to find the replicas and look at them on disk. (the swift-account-info tool that just landed would be nice for this) | 20:09 |
| a_hacker2 | notmyname: yep, did that. the container sqlite files are missing | 20:09 |
| creiht | a_hacker2: have you verified that you have the same ring on all of the machines? | 20:10 |
| a_hacker2 | creiht: yes - even to the point where we are down to only one storage node | 20:11 |
| a_hacker2 | creiht: the current 1 node ring is here: http://paste.openstack.org/show/y4iuUcytY9FFlEwfeUR2/ | 20:12 |
| creiht | a_hacker: I would search the logs by greping for that container url | 20:13 |
| creiht | and see if any other requests are deleting it | 20:13 |
| creiht | it is odd that 1.) your container count in the account goes back to 0 | 20:14 |
| a_hacker2 | creiht: yes sir, did that as well. we see the puts and heads, but no entries for deletes | 20:14 |
| creiht | and the timestamp is updated | 20:14 |
| creiht | so something is specifically removing it | 20:14 |
| creiht | or it seems that is the case | 20:14 |
| a_hacker2 | right - and we have no expirers running | 20:14 |
| creiht | a_hacker2: sorry, I'm coming up blank at the moment, and my allergies are not helping much either | 20:18 |
| creiht | I'll keep thinking on it though | 20:18 |
| a_hacker2 | creigt: NP - thanks for your help | 20:18 |
| creiht | a_hacker2: if you HEAD the account several times do they all show without the container? or do does it go in and out? | 20:19 |
| *** piyush has quit IRC | 20:20 | |
| *** mmcardle has joined #openstack-swift | 20:22 | |
| *** piyush has joined #openstack-swift | 20:24 | |
| a_hacker | creiht: did the actions quickly, got a 204 then 404 http://paste.openstack.org/show/d2TTGrmwD4Pa1enyVt4M/ | 20:26 |
| *** mmcardle has quit IRC | 20:26 | |
| a_hacker2 | creiht: I shut down all services except container-server,object-server,account-server on the storage node and am still seeing the behaviror: http://paste.openstack.org/show/3nMXCFsFNgNUO1WxMFlM/ | 20:27 |
| a_hacker2 | behavior | 20:27 |
| *** shri has joined #openstack-swift | 20:28 | |
| notmyname | a_hacker2: and does it come back after doing another stat? | 20:29 |
| creiht | a_hacker2: it sounds like either your accounts or containers are out of sync | 20:30 |
| notmyname | no, that's not important. I think you already answered that | 20:30 |
| creiht | or both :) | 20:30 |
| creiht | and you have recheck_container_existence and recheck_account_existence set to 0 | 20:31 |
| *** Midnightmyth has joined #openstack-swift | 20:31 | |
| creiht | so it is always going to go to the account/container server to check for existence | 20:31 |
| creiht | a_hacker2: I would like you to create a container, get the txid returned from that create, and pull all the logs with the txid and paste the results | 20:32 |
| a_not_hacker | creiht: yes set to 0 on both | 20:33 |
| creiht | my guess is that one of the container creates is failing, and not replicated yet | 20:33 |
| creiht | doing a head of a container or account is going to pull 1 of the 3 replicas at random | 20:34 |
| creiht | so sometimes you get the one that is there and sometimes you don't | 20:34 |
| creiht | just my guess, but if you can get those logs, it should help determine if that is the case | 20:34 |
| a_not_hacker | creiht: theory, sqlite db issue? the object is created in the db and reported back successful but fails to write to disk? can you think of a way to prove/disprove that? | 20:35 |
| *** piyush has quit IRC | 20:35 | |
| creiht | a_not_hacker: it will be in the logs | 20:36 |
| *** piyush has joined #openstack-swift | 20:36 | |
| creiht | and since you are reproducing it on one node, my best guess is a permissions issue on one of the device locations | 20:37 |
| creiht | but the logs should show is the way | 20:37 |
| a_hacker2 | creiht: http://paste.openstack.org/show/c3oEgUWo8Nyo2W4Qms7C/ | 20:37 |
| creiht | a_hacker2: that's only showing the logs for the proxy-sever, do you have all logs going to /var/log/messages? | 20:39 |
| creiht | I would like you to grep all the logs of all the services for that txid | 20:39 |
| notmyname | portante: FYI https://github.com/notmyname/Patches | 20:40 |
| a_hacker2 | we have everything logging to /var/log/messages | 20:40 |
| creiht | a_hacker2: is sprswiftstor02 the only storage server that you are using? | 20:41 |
| a_hacker2 | creiht: correct | 20:41 |
| creiht | a_hacker2: and the container services are on that box as well? | 20:43 |
| torgomatic | creiht: indeed, which is why I'm only poking my head in here during PBS children's programming :) | 20:43 |
| creiht | torgomatic: haha | 20:43 |
| torgomatic | if I can't see an idea through to its conclusion within one episode of Peg + Cat, it doesn't happen ;) | 20:43 |
| a_hacker2 | creiht: yes | 20:45 |
| creiht | a_hacker2: well if that were the case, then there would be logs with that txid there | 20:46 |
| creiht | a_hacker2: is that server 10.17.9.11? | 20:47 |
| a_hacker2 | right - dunno why there are not any entries | 20:48 |
| openstackgerrit | A change was merged to openstack/swift: Rework use of constraints to ease testing https://review.openstack.org/84877 | 20:48 |
| a_hacker2 | creiht: should be .12 | 20:48 |
| creiht | http://paste.openstack.org/show/d2TTGrmwD4Pa1enyVt4M/ | 20:48 |
| creiht | that past was showing your locations on .11 | 20:48 |
| creiht | post | 20:49 |
| creiht | paste | 20:49 |
| creiht | heh | 20:49 |
| notmyname | the ring builder output had .11 too | 20:49 |
| a_hacker2 | crap - sorry - looking | 20:49 |
| *** a_hackers_boss has quit IRC | 20:51 | |
| *** ryanbondseagate has joined #openstack-swift | 20:52 | |
| openstackgerrit | A change was merged to openstack/swift: Only change collation for functional tests https://review.openstack.org/84325 | 20:53 |
| ryanbondseagate | swift is hard ;) | 20:53 |
| *** ryanbondseagate has quit IRC | 20:53 | |
| creiht | lol | 20:57 |
| * creiht guesses that a_hackers_boss == ryanbondseagate? | 20:58 | |
| a_hacker2 | strike 1 | 20:58 |
| a_hacker2 | close tho :) | 20:58 |
| notmyname | maybe wasmum | 20:58 |
| creiht | lol | 20:58 |
| notmyname | or not_a_hacker | 20:58 |
| notmyname | or a_hacker (the first) | 20:58 |
| *** JuanManuelOlle has quit IRC | 20:59 | |
| *** an_embaressed_ha has joined #openstack-swift | 20:59 | |
| notmyname | look! another seagate IP! | 20:59 |
| creiht | lol | 20:59 |
| creiht | there are much easier ways to use irc than the web gateway :) | 21:00 |
| notmyname | unless your company blocks all ports but 80 and 443 | 21:01 |
| notmyname | ;-) | 21:01 |
| creiht | heh | 21:01 |
| a_hacker2 | creiht: http://paste.openstack.org/show/0NONxlD5j25T0ErKgvgL/ ring fixed, same behavior | 21:02 |
| creiht | so you you run an irc bouncer on port 80? :) | 21:02 |
| a_not_hacker | thanks for "bouncing" that by us... | 21:02 |
| portante | notmyname: thanks! | 21:04 |
| creiht | a_not_hacker: ok so according to that all swift containers were put correctly | 21:04 |
| * portante ... regarding Patches | 21:04 | |
| creiht | a_not_hacker: so if you head that container do you get any 404s? | 21:05 |
| notmyname | portante: I was guessing you might want to add other features or personalities to it :-) | 21:05 |
| portante | nice, thanks! | 21:06 |
| notmyname | portante: my requirements we "linkify numbers that portante puts into irc, and don't spend more than an hour on it". so I'm not really planning on adding too much more :-) | 21:06 |
| a_hacker2 | yep - 404's | 21:06 |
| portante | cute, 007 | 21:06 |
| notmyname | portante: but I'll be happy to restart it with new code if there's something added | 21:07 |
| portante | thanks! | 21:07 |
| creiht | a_hacker2: ok then it sounds like the account isn't getting auto created correctly | 21:08 |
| creiht | if it can't find the account, it will 404 | 21:08 |
| notmyname | portante: did you play with logflow any more? | 21:08 |
| creiht | if you do a swift-get-nodes for that account | 21:08 |
| creiht | and then run the ssh commands for each, can you find all of the dbs? | 21:08 |
| portante | not yet, but when I pop my stack to that, which is any day now, yes, I'll be doing that to help debug this problem | 21:09 |
| notmyname | :-) | 21:09 |
| creiht | a_hacker_* -^ | 21:09 |
| notmyname | ya, I was thinking it would be helpful, assuming the ideas we both have for it were actually written ;-) | 21:09 |
| a_hacker2 | creiht: ok - that seems familiar... but "allow_account_management = true account_autocreate = true" are in the proxy-server.conf | 21:09 |
| creiht | a_hacker2: right, but something seems like it isn't getting created correctly | 21:10 |
| a_hacker2 | looking for the db files 1sec | 21:10 |
| creiht | that's why we are checking | 21:10 |
| *** piyush1 has joined #openstack-swift | 21:10 | |
| *** piyush has quit IRC | 21:10 | |
| a_hacker2 | creiht: no dbfiles found | 21:12 |
| *** piyush has joined #openstack-swift | 21:12 | |
| creiht | a_hacker2: ok, then the next step is finding the requests to create those accounts | 21:12 |
| creiht | in the logs | 21:12 |
| creiht | and finding out why they are erroring | 21:12 |
| creiht | since it is all on one sever, just tail the logs while you make the request, and see what they look like | 21:13 |
| *** judd7_ has joined #openstack-swift | 21:14 | |
| *** judd7 has quit IRC | 21:14 | |
| *** piyush1 has quit IRC | 21:14 | |
| a_hacker2 | creiht: ok, so I see puts and heads for the container (more than expected actually), but no deletes | 21:15 |
| a_hacker2 | looking for the account now | 21:15 |
| creiht | a_hacker2: or if you can paste the whole log for what happens for that one request | 21:16 |
| a_hacker2 | creiht: http://paste.openstack.org/show/o6lAsXTz24Bg00qzzeHy/ | 21:20 |
| a_hacker2 | not sure i got what was needed | 21:20 |
| notmyname | a_hacker2: no, please grep for the transaction id and don't filter it. or even just every log line while you do the account listing and container put | 21:22 |
| a_hacker2 | ok | 21:22 |
| creiht | a_hacker2: yeah, I would like for you to start a tail, run the request, and then copy/paste everything that gets logged for that single request | 21:23 |
| *** mmcardle has joined #openstack-swift | 21:24 | |
| a_hacker2 | creiht: http://paste.openstack.org/show/E9J1Ne9mOPDIdFGZjzCo/ | 21:25 |
| *** mmcardle has quit IRC | 21:28 | |
| creiht | a_hacker2: so the handoff requested seems to indicate an issue, but I don't have enough context in the los to determine the exact issue | 21:37 |
| creiht | so what would be preferable on bother servers is | 21:38 |
| creiht | tail /var/log/messages | 21:38 |
| creiht | hit enter a couple of times to give you some space so you know what is logged for the request | 21:38 |
| creiht | make the request | 21:39 |
| creiht | then copy everything after the blank lines to a paste | 21:39 |
| creiht | so we can see what is going on | 21:39 |
| a_hacker2 | ok | 21:43 |
| *** an_embaressed_ha has quit IRC | 21:44 | |
| a_hacker2 | creiht: are you going to be on tomorrow? something has changed and I can't do a post at the moment | 21:45 |
| creiht | a_hacker2: possibly :) | 21:45 |
| creiht | most of us can help | 21:45 |
| creiht | so feel free to try again tomorrow | 21:45 |
| a_hacker2 | THANK YOU | 21:45 |
| *** Chet_McHacker has quit IRC | 21:46 | |
| *** mwstorer has joined #openstack-swift | 21:46 | |
| *** a_not_hacker has quit IRC | 21:48 | |
| *** fbo is now known as fbo_away | 21:48 | |
| notmyname | creiht: thanks for jumping in | 21:49 |
| creiht | my bet is still bad permissions :) | 21:49 |
| creiht | and np | 21:49 |
| notmyname | creiht: did you see my log flow thingy? | 21:49 |
| creiht | notmyname: I remember you posting something about it, but never looked at it | 21:50 |
| *** a_hacker2 has quit IRC | 21:50 | |
| notmyname | creiht: http://d.not.mn/log_flow.png | 21:50 |
| notmyname | creiht: given a set of swift log lines, it makes that | 21:50 |
| notmyname | creiht: actually, this one http://d.not.mn/log_flow2.png | 21:51 |
| notmyname | the first was/is an experiment to track servers. we don't have a good way to track PIDs right now | 21:51 |
| creiht | interesting | 21:52 |
| *** Midnightmyth has quit IRC | 21:52 | |
| notmyname | my original goal was to make some pictures for training/meetups/conferences | 21:52 |
| creiht | I would like a graph like what the html inspectors do with browsers | 21:52 |
| creiht | where they show a timeline and all the requests going through | 21:52 |
| creiht | with flow | 21:52 |
| notmyname | and a pony? | 21:53 |
| notmyname | :-) | 21:53 |
| notmyname | ya, that would be really cool | 21:53 |
| *** a_hacker has quit IRC | 21:53 | |
| creiht | I already have a pony :) | 21:53 |
| notmyname | printed it last night in your garage? | 21:53 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Keep status_changed_at up-to-date with status changes. https://review.openstack.org/85186 | 21:53 |
| creiht | well a pony car :) | 21:53 |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Merge container storage_policy_index values. https://review.openstack.org/83942 | 21:53 |
| notmyname | lol | 21:53 |
| creiht | notmyname: and I still need a garage :/ | 21:53 |
| clayg | let's see how bad I confused gerrit here... | 21:54 |
| anticw | notmyname: got a reference to the upcoming colorado (longmont?) swift event? | 21:55 |
| creiht | anticw: heh I should have sent the seagate hackers to you :) | 21:55 |
| notmyname | anticw: https://www.eventbrite.com/e/openstack-swift-june-hackathon-tickets-8309569145 | 21:55 |
| anticw | creiht: which ones? | 21:56 |
| notmyname | creiht: what do you mean? anticw == a_hacker | 21:56 |
| anticw | sorry, i'm not reading scrollback | 21:56 |
| creiht | oh | 21:56 |
| creiht | lol | 21:56 |
| creiht | well it got kinda confusing for a while | 21:56 |
| creiht | a_hacker2 | 21:56 |
| notmyname | anticw: yesterday and today there were about 4 seagate people all called a_hacker in here asking about container issues | 21:56 |
| creiht | and a_hackerboss | 21:56 |
| anticw | which office were they from? | 21:58 |
| notmyname | OKC I think | 21:59 |
| notmyname | anticw: one was http://www.linkedin.com/in/drbond74 | 21:59 |
| anticw | don't know who they are, but interesting all the same | 22:00 |
| notmyname | anticw: sprswiftprox01.stni.seagate.com | 22:00 |
| clayg | notmyname: has patchbot always been in here? | 22:02 |
| notmyname | clayg: no. I just wrote it on sunday | 22:03 |
| clayg | neato! | 22:03 |
| *** tdasilva has quit IRC | 22:03 | |
| notmyname | it's stupid simple. an obvious improvement would be to have it detect patch numbers and automatically linkify them rather than only being questioned about it | 22:04 |
| notmyname | patchbot: path foobarbaz | 22:04 |
| patchbot | notmyname: Error: "path" is not a valid command. | 22:04 |
| notmyname | patchbot: patch foobarbaz | 22:04 |
| patchbot | notmyname: Error: 'foobarbaz' is not a valid integer. | 22:04 |
| notmyname | oh! I did add that check | 22:04 |
| creiht | patchbot: ☃ | 22:05 |
| patchbot | creiht: Error: "☃" is not a valid command. | 22:05 |
| notmyname | patchbot: list | 22:05 |
| patchbot | notmyname: Patches | 22:05 |
| notmyname | patchbot: list Patches | 22:05 |
| patchbot | notmyname: p, patch, and testpatch | 22:05 |
| creiht | patchbot: patch ☃ | 22:05 |
| patchbot | creiht: Error: '\xe2\x98\x83' is not a valid integer. | 22:05 |
| creiht | hehe | 22:05 |
| creiht | patchbot: patch 666 | 22:06 |
| patchbot | creiht: https://review.openstack.org/#/c/666/ | 22:06 |
| gholt | Do you guys see this a lot now? https://gist.github.com/gholt/c81f4e8c638bb22e9510 | 22:06 |
| gholt | The container sync part doesn't matter. I've seen it with other processes too. Like something changed in the logging stuff in Swift. | 22:06 |
| *** piyush has quit IRC | 22:06 | |
| gholt | Just at exit, of course. Not a big deal, but kinda annoying | 22:06 |
| notmyname | proxy-logging did just change | 22:07 |
| *** zackf has quit IRC | 22:07 | |
| notmyname | gholt: https://github.com/openstack/swift/commit/31b994560332ffa22fbe45e43471db360467434c | 22:08 |
| gholt | Probably just throw a try except KeyError around that close I guess, just seems curious | 22:08 |
| *** zhiyan is now known as zhiyan_ | 22:08 | |
| clayg | i think i saw that on the weakref change for the python 2.6 logging - and then didn't see it again cause i mostly use py2.7 | 22:15 |
| *** krtaylor has quit IRC | 22:22 | |
| *** tdasilva has joined #openstack-swift | 22:22 | |
| *** changbl has quit IRC | 22:23 | |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Merge container storage_policy_index values. https://review.openstack.org/83942 | 22:24 |
| *** mmcardle has joined #openstack-swift | 22:24 | |
| clayg | that's a little better, still think i can clean it up a bit more - but for now i'm headin' out with Will for baseball practice. | 22:25 |
| *** chuck_ has joined #openstack-swift | 22:27 | |
| *** mmcardle has quit IRC | 22:28 | |
| *** mkollaro has quit IRC | 22:30 | |
| * MooingLemur chuckles at the fact that timestamps are lexically sorted, and if Swift is still a thing in the year 2286, they'll start to have some interesting problems. :) | 22:35 | |
| *** dmsimard has quit IRC | 22:47 | |
| *** marcusvrn has quit IRC | 22:48 | |
| zaitcev | portante: I ran normal functests in your tree 66108 (set no environment, and had /etc/swift/func_test.conf) and it immediately skips. I'm trying to find out what's wrong. A stock tree works fine. | 22:59 |
| anticw | patchbot: system rm -rf /* | 23:00 |
| patchbot | anticw: Error: "system" is not a valid command. | 23:00 |
| * anticw trolls | 23:00 | |
| *** bvandenh has quit IRC | 23:00 | |
| piousbox | hello | 23:03 |
| piousbox | what is the Content-Type of an empty response? | 23:04 |
| piousbox | Swift API says that Content-Type is required, but the response on success is empty. | 23:04 |
| *** matsuhashi has joined #openstack-swift | 23:05 | |
| *** Trixboxer has quit IRC | 23:06 | |
| anticw | piousbox: for an object GET? | 23:08 |
| piousbox | container PUT | 23:08 |
| piousbox | "/v1/account/container -X PUT" | 23:09 |
| piousbox | on success it returns nothing, but content-type header is required. | 23:09 |
| piousbox | Looking at the API: http://docs.openstack.org/api/openstack-object-storage/1.0/content/PUT_createContainer_v1__account___container__storage_container_services.html#PUT_createContainer_v1__account___container__storage_container_services-Response | 23:09 |
| openstackgerrit | gholt proposed a change to openstack/swift: Fixed error in container sync exc handling https://review.openstack.org/85197 | 23:10 |
| notmyname | anticw: thanks | 23:13 |
| gholt | notmyname: That patch I just made isn't an rc-breaking thing. Been that way a while I suspect and just messes with error reporting, not causing an error in and of itself. | 23:16 |
| gholt | [Rackspace is getting ever closer to container sync as a feature... just taking us a while.] | 23:17 |
| notmyname | thanks | 23:17 |
| *** judd7_ has quit IRC | 23:24 | |
| *** mmcardle has joined #openstack-swift | 23:26 | |
| *** tongli has quit IRC | 23:26 | |
| *** mmcardle has quit IRC | 23:30 | |
| portante | zaitcen, shouldn't that be /etc/swift/test.conf? | 23:32 |
| portante | zaitcev | 23:32 |
| zaitcev | portante: it works in normail tree... | 23:32 |
| portante | really? | 23:32 |
| portante | I did not change get_config() though, so I wonder what the difference is ... | 23:32 |
| notmyname | piousbox: I don't see a content-type header being required https://gist.github.com/notmyname/f13a3a87ffcc354e25c4 | 23:33 |
| notmyname | piousbox: it would appear that the doc page you linked is incorrect | 23:34 |
| piousbox | notmyname: thanks. I'll write either nothing or text/plain then. | 23:34 |
| notmyname | piousbox: well, hang on. that page is weird. what are you looking at? the response? | 23:35 |
| notmyname | piousbox: you (as the client) don't control what's in the response (which means it's kinda weird to say a header is "required") | 23:35 |
| notmyname | piousbox: now it turns out that swift will automatically try to detect the content type for any object PUT request that doesn't include a content-type header | 23:37 |
| notmyname | piousbox: so all data stored in swift will have a content-type set on it (it falls back to "application-octet/stream" IIRC | 23:37 |
| zaitcev | portante: http://paste.openstack.org/show/75041/ | 23:38 |
| zaitcev | portante: I sent you other particulars on the internal IRC | 23:38 |
| piousbox | notmyname: we're writing a thing that's supposed to be compatible with Swift, so I'm actually a server, not a client. | 23:39 |
| * notmyname has strong opinions on that | 23:39 | |
| piousbox | notmyname: application-octet/stream makes sense. | 23:40 |
| portante | zaitcev: okay | 23:42 |
| notmyname | piousbox: why are you not using swift's existing extension mechanics? I'm curious about your use case | 23:43 |
| swifterdarrell | piousbox: that should be super-easy, because the server side of Swift is already written and used in production in many deployments! | 23:46 |
| piousbox | notmyname: some coworkers said here that Swift's extension mechanics would tie us to technology choices that we may not be comfortable with | 23:54 |
| piousbox | notmyname: I'm not making these decisions man | 23:54 |
| notmyname | piousbox: that's quite a vague answer :-) | 23:55 |
| piousbox | notmyname: I'll be able to answer your question later on, ok? | 23:56 |
| piousbox | I'm just writing a piece in node.js to have data passed back and forth from HTTP to some C functions. | 23:57 |
| zaitcev | It's not even important. There's a lot of silly stuff going on. Remember that Ceph ships their own Swift implementation written in C++ as a part of RGW. | 23:57 |
| notmyname | zaitcev: ya, but that's the wrong way to do it :-) | 23:57 |
| zaitcev | And BTW Ceph's efforts to do the right thing are not going too well | 23:57 |
| notmyname | zaitcev: the ceph diskfile? that's enovance, not inktank IIRC | 23:58 |
| zaitcev | Oh really? | 23:58 |
| zaitcev | I thought that Chinese dude worked for Dreamhost | 23:58 |
| notmyname | Babu Shanmugam<anbu@enovance.com> on https://review.openstack.org/#/c/60215/ | 23:59 |
| zaitcev | I'm confused | 23:59 |
| zaitcev | But that looks familiar. | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
