| *** dosaboy has joined #openstack-swift | 00:23 | |
| *** rcernin has joined #openstack-swift | 00:48 | |
| *** gyee has quit IRC | 00:52 | |
| tdasilva | in tempurl we strip lose any other query param besides sig and expires, is that a security mechanism: https://github.com/openstack/swift/blob/master/swift/common/middleware/tempurl.py#L599 ? | 02:06 |
|---|---|---|
| tdasilva | it prevents a GET with version-id | 02:06 |
| tdasilva | no range requests either, right? | 02:11 |
| *** diablo_rojo has joined #openstack-swift | 03:30 | |
| *** psachin has joined #openstack-swift | 03:32 | |
| mattoliverau | well range is a header, so range requests probably work. But slo's also have problems, but that I think was by design. | 04:31 |
| mattoliverau | But it was one thing to go back and revisit | 04:31 |
| *** psachin has quit IRC | 04:48 | |
| timburke | so thinking about tempurl... i'm inclined to say that a today's notion of a tempurl for an object should only be useful for the *current* version -- i'm not entirely sure we should even include the version-id... | 04:54 |
| timburke | buuut... it *would* be nice to be able to have a new style of tempurl *for a particular version* -- maybe not so dissimilar from what we did by adding prefixes... | 04:55 |
| timburke | at least, that's my gut feeling. haven't actually tried to *implement it* yet, so take it with a grain of salt | 04:56 |
| timburke | probably OK to punt on tempurl-for-a-particular-version for now? i wonder what the s3api patch will let you do... | 04:57 |
| timburke | (since query params are included in canonical requests and all) | 04:57 |
| tdasilva | mattoliverau: yeah, got mixed up with range | 04:59 |
| tdasilva | timburke: the issue i'm seeing with tempurl+version-id is that it currently just ignores the version-id and sends back the latest version | 05:00 |
| timburke | 👍 i think i can live with that -- ?version-id is ignroed, just like all other query params | 05:01 |
| mattoliverau | yeah, I like it. if someone wants to give someone access to a version we could look at adding version to the HMAC somehow (like prefixes). | 05:04 |
| timburke | ?tempurl_version=xxx | 05:04 |
| tdasilva | ok, sounds good | 05:12 |
| mattoliverau | +1 | 05:22 |
| *** pcaruana has joined #openstack-swift | 06:02 | |
| *** pawan-gupta has joined #openstack-swift | 06:47 | |
| *** rcernin has quit IRC | 07:41 | |
| *** tesseract has joined #openstack-swift | 07:43 | |
| openstackgerrit | Merged openstack/swift master: py3: Make seamless reloads work https://review.opendev.org/698347 | 08:17 |
| *** FlorianFa has quit IRC | 08:49 | |
| *** FlorianFa has joined #openstack-swift | 09:01 | |
| *** FlorianFa has quit IRC | 09:09 | |
| *** mikecmpbll has joined #openstack-swift | 09:09 | |
| *** rcernin has joined #openstack-swift | 09:14 | |
| *** tkajinam has quit IRC | 09:19 | |
| *** FlorianFa has joined #openstack-swift | 09:22 | |
| *** rcernin has quit IRC | 09:53 | |
| *** rdejoux has joined #openstack-swift | 09:56 | |
| pawan-gupta | Hello friends, I have setup a openstack swift cluster, what I have noticed, if I make a s3v4 request, it takes more time (between 1-2 seconds extra) as compared to swift request, any specific reason behind this? Any help would be appreciated. thanks in advance | 10:42 |
| DHE | that's long enough that I would suspect something's failed and the proxy is trying to connect to it | 13:34 |
| DHE | aside: a machine with 30 hard drives and a single 1gig connection is disappointingly slow... | 14:28 |
| *** mikecmpbll has quit IRC | 15:36 | |
| clayg | i'm trying to duplicate lp bug #1856303 on 2.18 (rocky) since i'm not seeing the transfer-encoding: chunked repsonse to 404 on master | 15:46 |
| openstack | Launchpad bug 1856303 in OpenStack Object Storage (swift) "HEAD doesn't return Content-Length header due to catch_errors middleware" [Undecided,New] https://launchpad.net/bugs/1856303 | 15:46 |
| clayg | ... but I'm getting a signature does not match error - didn't we change something with the default region or something like that? | 15:47 |
| clayg | changing signature_version from s3v4 to s3 in .aws/config didn't help | 15:50 |
| *** baffle has quit IRC | 15:54 | |
| *** efried is now known as efried_afk | 16:12 | |
| clayg | location = us-east-1 in the filter:s3api config section 👍 | 16:14 |
| ormandj | heya folks, quick question (or not): what are people using for info such as container size for accounting/billing nowadays? looks like gnocchi is dead entirely | 16:20 |
| clayg | ormandj: swiftstack has a home rolled solution based roughly on the same design as slogging - https://github.com/notmyname/slogging | 16:25 |
| clayg | ormandj: for db in (find /srv/node*/*/accounts -name \*.db); do swift-account-info $db; done | 16:26 |
| *** gyee has joined #openstack-swift | 16:26 | |
| rledisez | ormandj: at OVH, we use a customer version of ceilometer that scan all sqlite db to extract info, then the ceilometer-collector push to mongodb through a custom driver implementation | 16:27 |
| rledisez | *custom version of ceilometer | 16:27 |
| *** rdejoux has quit IRC | 16:32 | |
| ormandj | copy that, thanks you two | 16:36 |
| *** tesseract has quit IRC | 16:53 | |
| clayg | rledisez: I think this one is going to bite whoever tries to deploy master -> 698563 | 16:58 |
| clayg | p 698563 | 16:58 |
| patchbot | https://review.opendev.org/#/c/698563/ - swift - WSGI server workers must drop_privledges - 2 patch sets | 16:58 |
| clayg | at least it's blocking my deploy 😁 | 16:58 |
| clayg | timburke: p 698799 seems like a good place to start - but I lost track of the gatekeeper quote_etags = false change? | 17:03 |
| patchbot | https://review.opendev.org/#/c/698799/ - swift - Add normalize_etag() helper function - 1 patch set | 17:03 |
| ormandj | clayg: rledisez: thank you. do you know what the long-term plans are for this accounting? if ceilometer wasn't just ingestion/notification, and had a data reporting/analytics api, swapping out backends would be easy, but as it is now, it seems like there's not really a great solution | 17:04 |
| ormandj | before we go write our own metrics gathering/reporting/api/etc goodies, i thought i'd check if there was actually something in the cards | 17:04 |
| rledisez | ormandj: you might be interested by this thread. it seems they will re-integrate mongodb | 17:05 |
| rledisez | http://lists.openstack.org/pipermail/openstack-discuss/2019-November/thread.html#10922 | 17:05 |
| ormandj | so idea is still to have all reporting done directly via the underlying storage vs. abstracting it so backend changes aren't so painful? | 17:08 |
| ormandj | or when ceilometer API is mentioned, is that the data-access API layer ? | 17:08 |
| ormandj | (from that thread) | 17:08 |
| ormandj | sorry - just trying to make sure I'm not offbase in my understanding :) | 17:08 |
| openstackgerrit | Clay Gerrard proposed openstack/python-swiftclient master: object versioning features https://review.opendev.org/691877 | 17:19 |
| clayg | ^ we'll see if that's any better on py3 | 17:19 |
| *** openstackgerrit has quit IRC | 17:29 | |
| *** mvkr has quit IRC | 17:34 | |
| timburke | good morning | 17:44 |
| *** openstackgerrit has joined #openstack-swift | 17:58 | |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: Let developers/operators add watchers to object audit https://review.opendev.org/212824 | 17:58 |
| ormandj | timburke: morning! | 18:47 |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: Let developers/operators add watchers to object audit https://review.opendev.org/212824 | 19:12 |
| *** efried_afk is now known as efried | 19:32 | |
| *** gmann is now known as gmann_afk | 20:13 | |
| clayg | i'm going to look at p 698799 | 20:23 |
| patchbot | https://review.opendev.org/#/c/698799/ - swift - Add normalize_etag() helper function - 1 patch set | 20:23 |
| clayg | see if the functest failure is anything 🤷♂️ | 20:24 |
| *** gyee has quit IRC | 20:35 | |
| *** gyee has joined #openstack-swift | 21:17 | |
| clayg | yeah that functest failure isn't anything, testFileThenContainerDelete I think is just maybe gunna exceed the retry limit very occasionally? | 21:18 |
| clayg | change looks good to me otherwise! | 21:18 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Test normalize_etag in swob https://review.opendev.org/699022 | 21:19 |
| *** mvkr has joined #openstack-swift | 22:38 | |
| *** pcaruana has quit IRC | 22:48 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!