| openstackgerrit | Tim Burke proposed openstack/swift master: WIP: s3api: Check whether versioning is enabled more https://review.opendev.org/722552 | 00:21 |
|---|---|---|
| timburke | lol! forgot about needing to address https://review.opendev.org/#/c/571021/ ... | 00:28 |
| patchbot | patch 571021 - swift - fix s3api functional tests - 1 patch set | 00:28 |
| *** gyee has quit IRC | 01:40 | |
| *** psachin has joined #openstack-swift | 03:21 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift master: WIP: PLaying around with a basic shard-cleanup tool https://review.opendev.org/722570 | 03:42 |
| mattoliverau | ^ that's just a POC and very much WIP. But kinda works. | 03:43 |
| mattoliverau | timburke: ^^ | 03:43 |
| mattoliverau | now I've been distracted enough, I need to go do some work :) | 03:43 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Add stable/ussuri upgrade job https://review.opendev.org/722571 | 04:14 |
| *** evrardjp has quit IRC | 04:35 | |
| *** evrardjp has joined #openstack-swift | 04:35 | |
| openstackgerrit | Merged openstack/swift stable/ussuri: func tests: Allow test_etag_quoter to be run multiple times https://review.opendev.org/722511 | 04:39 |
| mattoliverau | Now that's I've done the BIG HAMMER approach, next I'll try the misplaced objects gentle approach ;) | 04:40 |
| timburke | i thought you'd been distracted enough already :P | 04:48 |
| openstackgerrit | Matthew Oliver proposed openstack/swift master: WIP: PLaying around with a basic shard-cleanup tool https://review.opendev.org/722570 | 05:14 |
| mattoliverau | I have been :S but this is more interesting :) | 05:15 |
| mattoliverau | timburke: ^^ there is a the next version. Much smaller and less hammer like. | 05:15 |
| mattoliverau | just uses the root_container. | 05:15 |
| mattoliverau | Besides having a solution to clean up the shards in a root table will obviously be handy when we move to auto-sharding. | 05:19 |
| *** rcernin has quit IRC | 06:27 | |
| *** rcernin has joined #openstack-swift | 06:29 | |
| *** ccamacho has joined #openstack-swift | 06:37 | |
| openstackgerrit | Merged openstack/swift master: Add experimental job to upgrade from master https://review.opendev.org/722160 | 06:51 |
| *** rpittau|afk is now known as rpittau | 07:17 | |
| *** rcernin has quit IRC | 07:35 | |
| *** ccamacho has quit IRC | 07:47 | |
| *** ccamacho has joined #openstack-swift | 07:48 | |
| *** dtantsur|afk is now known as dtantsur | 08:17 | |
| *** tkajinam has quit IRC | 08:24 | |
| *** dtantsur is now known as dtantsur|brb | 09:15 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/python-swiftclient master: Fixed capability discovery endpoint hardcode https://review.opendev.org/722395 | 09:51 |
| *** rpittau is now known as rpittau|bbl | 10:12 | |
| *** psachin has quit IRC | 10:14 | |
| *** dtantsur|brb is now known as dtantsur | 11:14 | |
| *** rpittau|bbl is now known as rpittau | 12:13 | |
| viks____ | hi, does account/container ratelimiting work with s3api? anyone has any idea? | 13:43 |
| zaitcev | I don't see a connection. Ratelimit literally puts the request into sleep with eventlet.sleep(). The only potential issue is it ratelimits subrequests. | 15:49 |
| timburke | there's a bit of conflict with wanting to have ratelimit ahead of auth (so you don't burn resources figuring out whether the user is authorized just to find that they're ratelimited) but also after s3api (so you know what account/container you're rate-limiting) | 15:58 |
| timburke | following https://review.opendev.org/#/c/697535/ s3api should emit sane responses if ratelimit is right of s3api | 16:00 |
| patchbot | patch 697535 - swift - s3api: Better handle 498/429 responses (MERGED) - 1 patch set | 16:00 |
| timburke | so... maybe? i wouldn't say it's been widely tested | 16:06 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Add stable/ussuri upgrade job https://review.opendev.org/722571 | 16:17 |
| zaitcev | So, suggestions in etc/proxy-server.conf-sample are correct and up-to-date. | 16:26 |
| *** dtantsur is now known as dtantsur|afk | 16:29 | |
| *** evrardjp has quit IRC | 16:35 | |
| *** evrardjp has joined #openstack-swift | 16:35 | |
| *** rpittau is now known as rpittau|afk | 16:41 | |
| timburke | idk -- i'm not entirely convinced that `... s3api ratelimit auth ...` will work correctly :-/ | 16:48 |
| timburke | like, *maybe* it will if the account's clients are all-S3 or all-Swift, but i'm not sure that bimodal access will have unified rate limiting | 16:49 |
| timburke | also, with all the subrequests s3api makes, i'm not even sure that's a good idea -- i'm pretty sure ratelimit will consider the subrequests as having the same "weight" as a client request | 16:53 |
| *** rdejoux has quit IRC | 17:01 | |
| viks____ | zaitcev: timburke Thanks... To test, i had changed like this in my local cluster: | 17:40 |
| viks____ | in proxy server conf: | 17:40 |
| viks____ | ``` | 17:40 |
| viks____ | max_sleep_time_seconds = 2 | 17:40 |
| viks____ | container_ratelimit_10 = 1 | 17:40 |
| viks____ | ``` | 17:40 |
| viks____ | and my pipeline looks like: | 17:40 |
| viks____ | ``` | 17:40 |
| viks____ | pipeline = catch_errors gatekeeper healthcheck proxy-logging cache etag-quoter container_sync bulk formpost tempurl ratelimit crossdomain authtoken s3api s3token keystoneauth staticweb container-quotas account-quotas slo dlo versioned_writes symlink proxy-logging ceilometer proxy-server | 17:40 |
| viks____ | ``` | 17:40 |
| viks____ | I get 'HTTP/1.1 498 Rate Limited' for swift requests for some requests when did a parallel requests but not for parallel requests using awscli. I'll do more testing and get back with more details if possible on monday | 17:40 |
| timburke | yeah, with ratelimit left of s3api, it will no-op following https://review.opendev.org/#/c/540092/ (prior to that, it'd try to get account/container info and litter your logs with 404s or 400s) | 17:44 |
| patchbot | patch 540092 - swift - ratelimit: ignore requests with invalid API versions (MERGED) - 2 patch sets | 17:44 |
| timburke | well, no-op *for s3 requests* that is | 17:44 |
| viks____ | timburke: oh!! so should it be placed after `s3api s3token`.. what do you suggest? | 17:46 |
| viks____ | Also, for a bulk delete s3api request, i.e. https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteObjects.html, will this ratelimit apply? | 17:53 |
| *** mvkr has quit IRC | 18:49 | |
| *** mgagne has joined #openstack-swift | 19:52 | |
| *** mvkr has joined #openstack-swift | 20:16 | |
| -openstackstatus- NOTICE: The Zuul project gating service is reporting new patches in merge conflict erroneously due to a configuration error, fix in progress | 20:21 | |
| *** ccamacho has quit IRC | 20:51 | |
| timburke | viks____, i'm realizing that what i *really* want is to have ratelimit left of s3api, but install a callback in the environment that s3api can call once it actually knows what account/container the request is targetting. *maybe* that won't really be known until it hits the auth middleware though :-/ | 21:26 |
| timburke | speaking of s3api, i'm realizing that i want p 668064 ahead of p 722552 | 21:26 |
| patchbot | https://review.opendev.org/#/c/668064/ - swift - Give internal clients the option of getting contai... - 4 patch sets | 21:26 |
| patchbot | https://review.opendev.org/#/c/722552/ - swift - WIP: s3api: Check whether versioning is enabled more - 1 patch set | 21:26 |
| openstackgerrit | Tim Burke proposed openstack/swift master: s3api: Check whether versioning is enabled more https://review.opendev.org/722552 | 22:06 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Give internal clients the option of getting container HEADs from memcache https://review.opendev.org/668064 | 22:10 |
| openstackgerrit | Tim Burke proposed openstack/swift master: obj: _finalize_durable may succeed even when data file is missing https://review.opendev.org/722128 | 22:40 |
| -openstackstatus- NOTICE: the This Zuul outage was taken as an opportunity to perform an impromptu maintenance for changing our service deployment model; any merge failures received from Zuul between 19:40 and 20:20 UTC were likely in error and those changes should be rechecked; any patches uploaded between 20:55 and 22:45 UTC were missed entirely by Zuul and should also be rechecked to get fresh test results | 22:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!