openstackgerrit | Tim Burke proposed openstack/swift master: authors/changelog for 2.20.0 release https://review.openstack.org/625116 | 00:22 |
---|---|---|
notmyname | timburke: is that a rebase? | 00:27 |
timburke | nope | 00:27 |
notmyname | ah I see. was on the wrong compare in gerrit | 00:28 |
*** gyee has quit IRC | 00:39 | |
openstackgerrit | John Dickinson proposed openstack/swift master: add an upgrade job from rocky https://review.openstack.org/625376 | 00:53 |
openstackgerrit | Pete Zaitcev proposed openstack/swift master: py3: adapt the account server completely https://review.openstack.org/613505 | 00:59 |
*** nguyenhai_ has joined #openstack-swift | 02:24 | |
*** persia has joined #openstack-swift | 02:24 | |
*** nguyenhai has quit IRC | 02:27 | |
*** psachin has joined #openstack-swift | 04:54 | |
openstackgerrit | Merged openstack/swift master: Fix rolling upgrade job https://review.openstack.org/625361 | 05:10 |
openstackgerrit | Merged openstack/swift master: authors/changelog for 2.20.0 release https://review.openstack.org/625116 | 05:10 |
*** pcaruana has joined #openstack-swift | 08:00 | |
*** e0ne has joined #openstack-swift | 09:02 | |
*** e0ne has quit IRC | 09:06 | |
*** e0ne has joined #openstack-swift | 09:14 | |
*** e0ne has quit IRC | 09:19 | |
*** e0ne has joined #openstack-swift | 09:39 | |
*** e0ne has quit IRC | 09:41 | |
*** mikecmpbll has joined #openstack-swift | 09:43 | |
*** e0ne has joined #openstack-swift | 10:49 | |
*** e0ne has quit IRC | 10:53 | |
*** SkyRocknRoll has joined #openstack-swift | 11:12 | |
*** mikecmpbll has quit IRC | 11:22 | |
*** SkyRocknRoll has quit IRC | 11:31 | |
openstackgerrit | John Dickinson proposed openstack/swift master: add an upgrade job from rocky https://review.openstack.org/625376 | 15:50 |
*** patchbot has quit IRC | 16:46 | |
*** notmyname has quit IRC | 16:46 | |
*** notmyname has joined #openstack-swift | 16:47 | |
*** ChanServ sets mode: +v notmyname | 16:47 | |
*** patchbot has joined #openstack-swift | 16:47 | |
*** e0ne has joined #openstack-swift | 17:22 | |
*** e0ne has quit IRC | 17:38 | |
*** e0ne has joined #openstack-swift | 17:39 | |
*** e0ne has quit IRC | 17:40 | |
*** psachin has quit IRC | 17:43 | |
notmyname | current master can upgrade from previous versions just fine. https://review.openstack.org/#/c/625376/ /cc timburke | 18:32 |
patchbot | patch 625376 - swift - add an upgrade job from rocky - 2 patch sets | 18:32 |
*** spligak has joined #openstack-swift | 21:37 | |
spligak | We’re experiencing an interesting behavior with swift (mitaka release). I’m curious if anyone else has seen this. We’re running 12 processes of the swift-object-expirer daemon. Only the 8th process in the set is throwing errors. It hasn’t expired anything and has only received “400 bad request” responses from the various object servers. All of the other processes are merrily expiring things | 21:58 |
spligak | without error. | 21:58 |
spligak | I figured something might be wrong with the machine it was running on, so I stopped process 8 on its machine and brought it up on another. Exact same issue. It’s as if there’s something funky about the 8/12 sliver of work. | 21:59 |
spligak | Running out of ideas. Going to try reducing the number of processes to 10 and see if if (somehow) it was displeased with slicing things into 12 buckets? | 22:02 |
spligak | Oh my. This sounds familiar: https://github.com/openstack/swift/commit/57b632fbb52c84e02fefa307ada45081d86b23a6#diff-878167cc988d7e9129ad5185a2c381f9 | 22:10 |
spligak | That's exactly what we're seeing. The "Bad X-If-Delete-At header value" messages. | 22:11 |
notmyname | spligak: glad you found something. FWIW, you are able to upgrade directly from your current version deployed directly to a later version with no downtime. if you can afford to, I'd recommend waiting just a few days until I tag the 2.20.0 release. should be done by tuesday | 22:35 |
notmyname | the version of swift in the mitaka openstack cycle was 2.7.0 (from https://wiki.openstack.org/wiki/Swift/version_map) | 22:36 |
DHE | Regarding https://wiki.openstack.org/wiki/Swift/ideas/third-party-transfers ... would a potential middle-of-the-road option be the ability to upload an object by specifying a remote URL to fetch? plus headers that are needed (eg: X-RemoteHeader-Auth-Token) in the GET or such... | 23:26 |
DHE | it reminded me of a hack to the FTP protocol, sometimes known as FXP, where two FTP servers connect to each other for the bulk data transfer even though the client is the one actually connected to both... | 23:27 |
spligak | notmyname: thank you! | 23:30 |
notmyname | DHE: I think the answer to that idea wiki page is to use 1space | 23:32 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!