| *** baojg has quit IRC | 00:00 | |
| *** baojg has joined #openstack-swift | 00:00 | |
| *** baojg has quit IRC | 00:00 | |
| *** baojg has joined #openstack-swift | 00:01 | |
| *** baojg has quit IRC | 00:01 | |
| *** baojg has joined #openstack-swift | 00:02 | |
| *** baojg has quit IRC | 00:02 | |
| *** baojg has joined #openstack-swift | 00:02 | |
| *** baojg has quit IRC | 00:03 | |
| *** baojg has joined #openstack-swift | 00:03 | |
| *** baojg has quit IRC | 00:04 | |
| *** baojg has joined #openstack-swift | 00:04 | |
| *** baojg has quit IRC | 00:04 | |
| *** baojg has joined #openstack-swift | 00:05 | |
| *** baojg has quit IRC | 00:05 | |
| *** baojg has joined #openstack-swift | 00:06 | |
| *** baojg has quit IRC | 00:06 | |
| *** baojg has joined #openstack-swift | 00:06 | |
| *** baojg has quit IRC | 00:07 | |
| *** baojg has joined #openstack-swift | 00:07 | |
| *** baojg has quit IRC | 00:08 | |
| *** baojg has joined #openstack-swift | 00:08 | |
| *** baojg has quit IRC | 00:08 | |
| *** baojg has joined #openstack-swift | 00:09 | |
| *** baojg has quit IRC | 00:09 | |
| *** baojg has joined #openstack-swift | 00:10 | |
| *** baojg has quit IRC | 00:10 | |
| *** baojg has joined #openstack-swift | 00:10 | |
| *** baojg has quit IRC | 00:11 | |
| *** baojg has joined #openstack-swift | 00:12 | |
| *** baojg has quit IRC | 00:12 | |
| *** baojg has joined #openstack-swift | 00:13 | |
| *** baojg has quit IRC | 00:13 | |
| *** baojg has joined #openstack-swift | 00:14 | |
| *** baojg has quit IRC | 00:14 | |
| *** baojg has joined #openstack-swift | 00:14 | |
| *** baojg has quit IRC | 00:15 | |
| *** baojg has joined #openstack-swift | 00:15 | |
| *** baojg has quit IRC | 00:16 | |
| *** baojg has joined #openstack-swift | 00:16 | |
| *** baojg has quit IRC | 00:16 | |
| *** baojg has joined #openstack-swift | 00:17 | |
| *** baojg has quit IRC | 00:17 | |
| *** baojg has joined #openstack-swift | 00:18 | |
| *** baojg has quit IRC | 00:18 | |
| *** baojg has joined #openstack-swift | 00:18 | |
| *** baojg has quit IRC | 00:19 | |
| *** sasregulus has quit IRC | 02:05 | |
| *** baojg has joined #openstack-swift | 02:19 | |
| *** baojg has quit IRC | 02:24 | |
| *** psachin has joined #openstack-swift | 03:25 | |
| *** baojg has joined #openstack-swift | 03:44 | |
| *** viks___ has joined #openstack-swift | 04:34 | |
| *** new_student1411 has joined #openstack-swift | 05:02 | |
| *** rcernin has quit IRC | 05:27 | |
| *** manuvakery has joined #openstack-swift | 05:28 | |
| *** rcernin has joined #openstack-swift | 06:44 | |
| *** fungi has quit IRC | 06:50 | |
| *** mugsie has quit IRC | 06:50 | |
| *** niceplace_ has quit IRC | 06:50 | |
| *** csmart has quit IRC | 06:50 | |
| *** hoonetorg has quit IRC | 07:02 | |
| *** aluria has joined #openstack-swift | 07:04 | |
| *** niceplace_ has joined #openstack-swift | 07:16 | |
| *** csmart has joined #openstack-swift | 07:16 | |
| *** fungi has joined #openstack-swift | 07:16 | |
| *** mugsie has joined #openstack-swift | 07:16 | |
| *** openstackgerrit has quit IRC | 07:18 | |
| *** hoonetorg has joined #openstack-swift | 07:19 | |
| *** ccamacho has joined #openstack-swift | 07:38 | |
| *** niceplace_ has quit IRC | 08:26 | |
| *** csmart has quit IRC | 08:26 | |
| *** fungi has quit IRC | 08:27 | |
| *** mugsie has quit IRC | 08:27 | |
| *** persia has joined #openstack-swift | 08:35 | |
| *** niceplace has joined #openstack-swift | 08:36 | |
| *** fungi has joined #openstack-swift | 08:37 | |
| *** mugsie has joined #openstack-swift | 08:37 | |
| *** csmart has joined #openstack-swift | 08:42 | |
| *** e0ne has joined #openstack-swift | 08:45 | |
| *** rdejoux has joined #openstack-swift | 08:52 | |
| *** rcernin has quit IRC | 08:55 | |
| *** e0ne has quit IRC | 09:00 | |
| *** e0ne has joined #openstack-swift | 09:12 | |
| *** e0ne has quit IRC | 09:25 | |
| *** baojg has quit IRC | 11:19 | |
| *** tesseract has joined #openstack-swift | 11:20 | |
| *** baojg has joined #openstack-swift | 11:25 | |
| *** pcaruana has joined #openstack-swift | 11:33 | |
| *** psachin has quit IRC | 11:36 | |
| *** baojg has quit IRC | 11:45 | |
| *** baojg has joined #openstack-swift | 11:45 | |
| *** baojg has quit IRC | 11:46 | |
| *** baojg has joined #openstack-swift | 11:46 | |
| *** baojg has quit IRC | 11:47 | |
| *** baojg has joined #openstack-swift | 11:47 | |
| *** baojg has quit IRC | 11:47 | |
| *** baojg has joined #openstack-swift | 11:48 | |
| *** baojg has quit IRC | 11:48 | |
| *** tesseract has quit IRC | 11:48 | |
| *** baojg has joined #openstack-swift | 11:49 | |
| *** baojg has quit IRC | 11:49 | |
| *** baojg has joined #openstack-swift | 11:49 | |
| *** baojg has quit IRC | 11:50 | |
| *** baojg has joined #openstack-swift | 11:50 | |
| *** baojg has quit IRC | 11:51 | |
| *** baojg has joined #openstack-swift | 11:51 | |
| *** tesseract has joined #openstack-swift | 11:51 | |
| *** baojg has quit IRC | 11:51 | |
| *** baojg has joined #openstack-swift | 11:52 | |
| *** baojg has quit IRC | 11:52 | |
| *** baojg has joined #openstack-swift | 11:53 | |
| *** baojg has quit IRC | 11:53 | |
| *** baojg has joined #openstack-swift | 11:53 | |
| *** baojg has quit IRC | 11:54 | |
| *** baojg has joined #openstack-swift | 11:54 | |
| *** baojg has quit IRC | 11:55 | |
| *** baojg has joined #openstack-swift | 11:55 | |
| *** baojg has quit IRC | 11:55 | |
| *** baojg has joined #openstack-swift | 11:56 | |
| *** baojg has quit IRC | 11:56 | |
| *** baojg has joined #openstack-swift | 11:56 | |
| *** baojg has quit IRC | 11:57 | |
| *** baojg has joined #openstack-swift | 11:57 | |
| *** baojg has quit IRC | 11:58 | |
| *** baojg has joined #openstack-swift | 11:58 | |
| *** baojg has quit IRC | 11:59 | |
| *** baojg has joined #openstack-swift | 11:59 | |
| *** baojg has quit IRC | 11:59 | |
| *** baojg has joined #openstack-swift | 12:00 | |
| *** baojg has quit IRC | 12:00 | |
| *** baojg has joined #openstack-swift | 12:00 | |
| *** baojg has quit IRC | 12:01 | |
| *** baojg has joined #openstack-swift | 12:01 | |
| *** baojg has quit IRC | 12:02 | |
| *** baojg has joined #openstack-swift | 14:02 | |
| *** baojg has quit IRC | 14:07 | |
| *** manuvakery has quit IRC | 14:19 | |
| *** new_student1411 has quit IRC | 14:19 | |
| *** baojg has joined #openstack-swift | 14:20 | |
| *** tesseract has quit IRC | 14:28 | |
| *** tesseract has joined #openstack-swift | 14:30 | |
| *** irclogbot_3 has quit IRC | 14:40 | |
| *** irclogbot_1 has joined #openstack-swift | 14:51 | |
| *** sasregulus has joined #openstack-swift | 14:58 | |
| *** tesseract has quit IRC | 15:02 | |
| *** tesseract has joined #openstack-swift | 15:03 | |
| *** gyee has joined #openstack-swift | 15:31 | |
| timburke | DHE, it goes by timestamp. so if the x-timestamps say it went POST-then-PUT, the metadata from the PUT wins. if the timestamps say it went PUT-then-POST, the metadata from the POST wins (but of course, the new *data* will be the same either way) | 16:08 |
|---|---|---|
| DHE | timburke: I'm looking at making a custom diskfile implementation, but want to do it right. worried about a race between multiple object servers completing a PUT racing against an incoming POST | 16:22 |
| DHE | my plan is to avoid the one-directory-per-object current configuration, handling multiple uploads and tombstones in different ways | 16:23 |
| zigo | timburke: Building Swift for Python 3 in Debian Sid, I get a single error: http://paste.openstack.org/show/754397/ | 16:40 |
| zigo | Any idea what's that one about? :) | 16:40 |
| zigo | My /tmp is a normal ext4 filesystem, so it should be fine to use xattr... | 16:42 |
| zigo | To me, it looks like the if errno.errorcode.get(e.errno) in ('ENOSPC', 'ENOTSUP', 'EOPNOTSUPP'): is missing some values, somehow ... | 16:44 |
| * zigo is trying to print the value of the exception... | 16:50 | |
| timburke | zigo, here, i guess? https://github.com/torvalds/linux/blob/fec88ab0af9706b2201e5daf377c5031c62d11f7/fs/xattr.c#L424-L426 | 17:09 |
| timburke | hmm... | 17:09 |
| zigo | So, ERANGE ? | 17:10 |
| * zigo is still not finished to prepare a test env | 17:10 | |
| timburke | at least, that's what *my* errno.errorcode[34] says | 17:11 |
| *** e0ne has joined #openstack-swift | 17:12 | |
| zigo | Will try. | 17:13 |
| *** baojg has quit IRC | 17:14 | |
| *** baojg has joined #openstack-swift | 17:15 | |
| *** baojg has quit IRC | 17:15 | |
| *** baojg has joined #openstack-swift | 17:15 | |
| timburke | so i guess the xattr *name* is too long... but 22 chars doesn't seem *so* long? i wonder where/how XATTR_NAME_MAX gets defined... | 17:16 |
| *** baojg has quit IRC | 17:16 | |
| timburke | yeah, 255 seems more reasonable: https://github.com/torvalds/linux/blob/851ca779d110f694b5d078bc4af06d3ad37169e8/include/uapi/linux/limits.h#L15 | 17:16 |
| *** baojg has joined #openstack-swift | 17:16 | |
| *** baojg has quit IRC | 17:17 | |
| *** baojg has joined #openstack-swift | 17:17 | |
| *** baojg has quit IRC | 17:17 | |
| *** baojg has joined #openstack-swift | 17:18 | |
| *** baojg has quit IRC | 17:18 | |
| *** baojg has joined #openstack-swift | 17:19 | |
| *** baojg has quit IRC | 17:19 | |
| *** baojg has joined #openstack-swift | 17:19 | |
| timburke | the ext4 code for it looks similar: https://github.com/torvalds/linux/blob/master/fs/ext4/xattr.c#L2317-L2318 | 17:19 |
| zigo | timburke: This fixes it! :) | 17:20 |
| zigo | timburke: I'll open the patch review soon. | 17:20 |
| *** baojg has quit IRC | 17:20 | |
| *** baojg has joined #openstack-swift | 17:20 | |
| timburke | but it's so weird! what's the xattr name limit for you?? | 17:20 |
| *** baojg has quit IRC | 17:21 | |
| *** baojg has joined #openstack-swift | 17:21 | |
| zigo | Oh, what is swift-container-deleter for? | 17:21 |
| *** baojg has quit IRC | 17:21 | |
| *** baojg has joined #openstack-swift | 17:22 | |
| *** baojg has quit IRC | 17:22 | |
| *** baojg has joined #openstack-swift | 17:23 | |
| *** baojg has quit IRC | 17:23 | |
| *** baojg has joined #openstack-swift | 17:23 | |
| *** baojg has quit IRC | 17:24 | |
| *** baojg has joined #openstack-swift | 17:24 | |
| *** baojg has quit IRC | 17:25 | |
| *** altlogbot_3 has quit IRC | 17:25 | |
| *** baojg has joined #openstack-swift | 17:25 | |
| *** baojg has quit IRC | 17:25 | |
| *** irclogbot_1 has quit IRC | 17:26 | |
| *** baojg has joined #openstack-swift | 17:26 | |
| *** baojg has quit IRC | 17:26 | |
| *** irclogbot_0 has joined #openstack-swift | 17:29 | |
| timburke | there are a few ways to clean up data in swift: you delete a bunch of names one by one, you can send a list of names to the bulk-delete api if that's enabled, or you can mark an entire account as deleted and have the account-reaper clean it up in the background | 17:31 |
| *** irclogbot_0 has quit IRC | 17:32 | |
| zigo | timburke: So, swift-container-deleter is a service that will do this in the background? | 17:33 |
| timburke | now we've got a fourth way: operators (at the moment; i'm hopeful about turning this into a client-facing api) can delete portions of a single container's namespace. it does listings against the container and queues up deletes for the expirer to process async | 17:33 |
| timburke | not a service; command line tool. but yeah | 17:34 |
| timburke | i'd like to expand on it in two major ways: first, expose it as a client-facing api. second, make it so you can "reap" a single container -- mark it such that subsequent requests to it will 410, and have a background daemon (replication, maybe?) queue up expirer work to clear it out. once all the data's gone (and some additional amount of time has passed, to ensure async pendings have landed) actually delete the DB | 17:37 |
| timburke | but this was a good first step | 17:37 |
| *** irclogbot_1 has joined #openstack-swift | 17:39 | |
| *** altlogbot_1 has joined #openstack-swift | 17:41 | |
| *** tesseract has quit IRC | 17:45 | |
| *** openstackgerrit has joined #openstack-swift | 17:54 | |
| openstackgerrit | Thomas Goirand proposed openstack/swift master: Fix test_parse_get_node_args https://review.opendev.org/670894 | 17:54 |
| zigo | There you go! | 17:58 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Authors/changelog for 2.22.0 https://review.opendev.org/668990 | 18:13 |
| clayg | timburke: Thiago pointed out that the character check in hardlinks is going to make it difficult to do a hardlink to a manifest: https://review.opendev.org/#/c/633094/12/swift/common/middleware/symlink.py@266 | 18:38 |
| patchbot | patch 633094 - swift - Allow "harder" symlinks - 12 patch sets | 18:38 |
| clayg | In fixing that up I had to allow symlink to peak into 'x-object-sysmeta-slo-etag' (since it's right of slo) - but also remove the double quote from the dis-allowed chars 🤔 | 18:39 |
| clayg | would it be better to maybe ... strip " and *then* check for disallowed chars? | 18:39 |
| clayg | I can't think of a good way to avoid leaking the slo sysmeta name into symlink | 18:39 |
| *** rdejoux has quit IRC | 18:49 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Allow "harder" symlinks https://review.opendev.org/633094 | 18:56 |
| zigo | timburke: I haven't tested much, but so far, Swift on Pytohn 3 worked for me in Debian. I was able to create a container and an object in it... | 18:58 |
| timburke | \O/ | 18:59 |
| timburke | always good when the sniff-tests pass :-) | 18:59 |
| zigo | I'm uploading to Debian experimental, so that we can process other packages first (swauth and others...) | 18:59 |
| zigo | timburke: You haven't released since Stein, right? | 18:59 |
| zigo | timburke: Is an intermediary release planned before Train? | 19:00 |
| timburke | clayg, yeah, that kinda sucks... and it'll need the s3 sysmeta variant, too... | 19:00 |
| timburke | zigo, no, we haven't released since stein. yes, we're planning to. real soon, in fact; i was waiting for a few more py3 patches to land, but now i think i'm good to move forward on https://review.opendev.org/#/c/668990/ | 19:01 |
| patchbot | patch 668990 - swift - Authors/changelog for 2.22.0 - 4 patch sets | 19:01 |
| zigo | Good to know! | 19:02 |
| clayg | idk, this isn't great - now versoined writes needs to know about slo too so that it can create hard links to manifests | 19:21 |
| clayg | maybe hardlinks to manifests should use the manifest's objects etag... it's in the listing and you can get it with `multipart-manifest=get` | 19:21 |
| clayg | as it turns out it's also still in the put_resp when versioned writes gets ahold of it - because slo hasn't clobbered it with the calculated value it stuffed in sysmeta | 19:22 |
| *** baojg has joined #openstack-swift | 19:27 | |
| *** baojg has quit IRC | 19:32 | |
| clayg | well, that may not be quite right - SLO put response encodes a lot of detail in the response body at the end of the heartbeat | 19:33 |
| clayg | the actual PUT resp of the manifest comes through versioned writes mw, but isn't retruned to the client... | 19:33 |
| clayg | I think from a client perspective it's better to let them provide the manifest etag tho... so I guess versioned_writes will just have to learn how to spot a SLO manifest PUT | 19:34 |
| *** e0ne has quit IRC | 19:52 | |
| *** e0ne has joined #openstack-swift | 19:53 | |
| clayg | maybe I'll allow .... both 😬 | 19:58 |
| clayg | https://media0.giphy.com/media/146heXDX89mUgw/source.gif | 19:58 |
| *** altlogbot_1 has quit IRC | 20:10 | |
| *** altlogbot_0 has joined #openstack-swift | 20:14 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Allow "harder" symlinks https://review.opendev.org/633094 | 20:20 |
| *** pcaruana has quit IRC | 20:29 | |
| clayg | jesus, now I'm parsing slo_etag of of container listing etag feilds so that I have it handy for when I do stack pop 😞 | 20:36 |
| *** BjoernT has joined #openstack-swift | 20:37 | |
| *** BjoernT_ has joined #openstack-swift | 20:41 | |
| *** BjoernT has quit IRC | 20:42 | |
| *** persia has left #openstack-swift | 20:47 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: WIP: symlink-backed versioned_writes https://review.opendev.org/633857 | 20:53 |
| *** e0ne has quit IRC | 21:03 | |
| *** e0ne has joined #openstack-swift | 21:19 | |
| *** e0ne has quit IRC | 21:20 | |
| *** e0ne has joined #openstack-swift | 21:22 | |
| *** e0ne has quit IRC | 21:24 | |
| *** e0ne has joined #openstack-swift | 21:24 | |
| *** baojg has joined #openstack-swift | 21:28 | |
| *** baojg has quit IRC | 21:32 | |
| *** e0ne has quit IRC | 21:39 | |
| *** BjoernT_ has quit IRC | 22:07 | |
| timburke | clayg, what would you expect the py2/py3 option to look like in vsaio? `export PYTHON=python2 # or python3`? `export USE_PYTHON3=false # or true`? | 22:31 |
| clayg | Use py3 wfm - probably less work to support? | 22:34 |
| clayg | Like if someday something only works in py3 we don’t have to parse the python option 🤔 | 22:34 |
| mattoliverau | morning | 22:53 |
| *** tkajinam has joined #openstack-swift | 22:59 | |
| *** rcernin has joined #openstack-swift | 23:29 | |
| *** baojg has joined #openstack-swift | 23:29 | |
| *** e0ne has joined #openstack-swift | 23:31 | |
| *** baojg has quit IRC | 23:33 | |
| *** e0ne has quit IRC | 23:36 | |
| *** e0ne has joined #openstack-swift | 23:38 | |
| *** e0ne has quit IRC | 23:42 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
