| *** rcernin has quit IRC | 00:27 | |
| *** rcernin has joined #openstack-swift | 00:29 | |
| *** two_tired has joined #openstack-swift | 00:46 | |
| timburke | mattoliverau: on the kmip logging -- iirc, it's just a matter of turning on debug logging for the proxy-server. i think i maybe had it up in the [DEFAULT] section? then when we attach the middleware's handlers at https://github.com/openstack/swift/blob/2.19.0/swift/common/middleware/crypto/kmip_keymaster.py#L119-L121 ... you see things like | 01:03 |
|---|---|---|
| timburke | Oct 4 18:24:44 centos74node1 kmip.services.server.kmip_protocol: KMIPProtocol.read: 42007b010000011042007a0100000048420069010000002042006a0200000004000000010000000042006b020000000400000002000000004200920900000008000000005bb65aec42000d0200000004000000010000000042000f01000000b842005c05000000040000000a0000000042007f0500000004000000000000000042007c0100000090420057050000000400000002000000004200940700000001310000000000000042008f01000000684 | 01:03 |
| timburke | 2004001000000604200420500000004000000010000000042004501000000284200430800000020d24274b2a4e37a10987efc291313c4ed80b9519251eac772fed96d76408c14164200280500000004000000030000000042002a02000000040000010000000000 | 01:03 |
| mattoliverau | timburke: oh, that's what I was trying. But wasn't seeing anything. Maybe I've got something wrong | 01:04 |
| *** _david_sohonet has quit IRC | 01:04 | |
| mattoliverau | ie, turning debug level logging on in the proxy | 01:04 |
| timburke | (it's a junk key anyway, so i don't mind you knowing that the "d24274b2a4e37a10987efc291313c4ed80b9519251eac772fed96d76408c1416" is my hex-encoded key) | 01:05 |
| timburke | hmm.... | 01:05 |
| *** two_tired has quit IRC | 01:06 | |
| mattoliverau | timburke: are you using a proper kmip implementation. Maybe the demo env acoles setup is slightly different. | 01:08 |
| timburke | huh. so i don't see it on my all-in-one... but when i was testing with what i plan to ship to customers, i do... | 01:14 |
| timburke | it all comes down to https://github.com/OpenKMIP/PyKMIP/blob/v0.8.0/kmip/services/server/kmip_protocol.py#L50-L51 | 01:15 |
| timburke | now i want to figure out what's different, but i ought to head home... | 01:15 |
| mattoliverau | from the `kmip.services.server.kmip_protocol` it seems your connecting the kmip servers log, not nessesarily the clients. | 01:16 |
| mattoliverau | maybe your production implementation your running a server inline as well as the client? | 01:16 |
| mattoliverau | timburke: but yeah, head home. spend time with the fam :) | 01:17 |
| timburke | nah, the client's definitely using the same kmip_protocol module to read/write (as it should, i think): https://github.com/OpenKMIP/PyKMIP/blob/v0.8.0/kmip/services/kmip_client.py#L53 | 01:19 |
| * timburke heads out before i go too far down this rabbit hole | 01:19 | |
| timburke | i'll let you know what i find out tomorrow | 01:19 |
| mattoliverau | lol, sorry :) | 01:22 |
| *** aluria has quit IRC | 01:29 | |
| *** gyee has quit IRC | 02:01 | |
| *** itlinux has joined #openstack-swift | 02:05 | |
| *** dosaboy has quit IRC | 02:35 | |
| *** rcernin has quit IRC | 04:42 | |
| *** rcernin has joined #openstack-swift | 04:46 | |
| *** kukacz has quit IRC | 04:52 | |
| *** guimaluf has quit IRC | 04:52 | |
| *** openstackgerrit has quit IRC | 04:52 | |
| *** d0ugal has quit IRC | 04:52 | |
| *** kukacz has joined #openstack-swift | 04:57 | |
| *** guimaluf has joined #openstack-swift | 04:57 | |
| *** openstackgerrit has joined #openstack-swift | 04:57 | |
| *** d0ugal has joined #openstack-swift | 04:57 | |
| *** e0ne has joined #openstack-swift | 05:34 | |
| *** e0ne has quit IRC | 05:50 | |
| *** e0ne has joined #openstack-swift | 05:53 | |
| *** e0ne has quit IRC | 06:12 | |
| *** alecuyer has quit IRC | 06:48 | |
| *** alecuyer has joined #openstack-swift | 06:52 | |
| *** pcaruana has joined #openstack-swift | 06:57 | |
| *** fatema__ has joined #openstack-swift | 06:59 | |
| *** rcernin has quit IRC | 07:04 | |
| *** dosaboy has joined #openstack-swift | 07:38 | |
| *** aluria has joined #openstack-swift | 07:47 | |
| *** e0ne has joined #openstack-swift | 07:49 | |
| *** mikecmpbll has joined #openstack-swift | 07:59 | |
| *** gkadam has joined #openstack-swift | 09:43 | |
| *** sheel has joined #openstack-swift | 10:05 | |
| *** itlinux has quit IRC | 10:24 | |
| *** fatema__ has quit IRC | 11:24 | |
| *** e0ne has quit IRC | 11:54 | |
| *** e0ne has joined #openstack-swift | 12:29 | |
| *** sheel has quit IRC | 12:34 | |
| *** dosaboy has quit IRC | 12:35 | |
| *** dosaboy has joined #openstack-swift | 12:36 | |
| *** dosaboy has quit IRC | 12:36 | |
| *** dosaboy has joined #openstack-swift | 12:56 | |
| *** two_tired has joined #openstack-swift | 13:48 | |
| *** two_tired has quit IRC | 14:38 | |
| *** e0ne has quit IRC | 14:38 | |
| *** gyee has joined #openstack-swift | 15:24 | |
| *** pcaruana has quit IRC | 15:39 | |
| *** cwright_ has quit IRC | 15:46 | |
| *** cwright has joined #openstack-swift | 15:50 | |
| *** e0ne has joined #openstack-swift | 15:50 | |
| *** e0ne has quit IRC | 15:55 | |
| *** mrjk_ has quit IRC | 15:56 | |
| *** mrjk_ has joined #openstack-swift | 15:57 | |
| *** mrjk_ has quit IRC | 15:58 | |
| *** mrjk_ has joined #openstack-swift | 15:58 | |
| *** mrjk_ has quit IRC | 16:00 | |
| *** mikecmpbll has quit IRC | 16:08 | |
| *** itlinux has joined #openstack-swift | 16:12 | |
| *** ianychoi_ is now known as ianychoi | 16:19 | |
| timburke | good morning | 16:20 |
| *** _david_sohonet has joined #openstack-swift | 16:20 | |
| *** gkadam has quit IRC | 16:21 | |
| notmyname | good morning | 16:33 |
| cwright | notmyname: good morning | 16:36 |
| cwright | notmyname: Just following back up on something from a few weeks ago, have you had a chance to talk to the team there about open sourcing the Prometheus Swift alerts that SwiftStack is currently using? | 16:36 |
| *** pcaruana has joined #openstack-swift | 16:42 | |
| notmyname | cwright: I'll check | 16:45 |
| *** pcaruana has quit IRC | 16:50 | |
| *** mikecmpbll has joined #openstack-swift | 16:58 | |
| cwright | notmyname: thanks! | 17:24 |
| openstackgerrit | Tim Burke proposed openstack/swift master: internal_client: Require that request_tries be positive https://review.openstack.org/608330 | 17:51 |
| timburke | request_tries, request_retries... w/e | 17:52 |
| timburke | i think i got confused because swiftclient uses retries and *does* let you say 0... but internal_client didn't like it when i said to try zero times... | 17:53 |
| *** e0ne has joined #openstack-swift | 19:38 | |
| *** e0ne has quit IRC | 19:43 | |
| zaitcev | okay | 19:49 |
| zaitcev | tried tox -e lower-constraints | 19:49 |
| zaitcev | result is a nosetests ballooning to VIRT 9.862g RES 7.825g | 19:54 |
| zaitcev | That sounds like no constraints at all lol | 19:54 |
| *** e0ne has joined #openstack-swift | 19:55 | |
| notmyname | lol | 19:55 |
| *** e0ne has quit IRC | 19:56 | |
| zaitcev | https://review.openstack.org/#/c/581905/ failed check on lower-constraints, so I thought I'd take a look. Even with FF quit, not enough RAM | 19:57 |
| patchbot | patch 581905 - swift - py3: Adapt db.py - 5 patch sets | 19:57 |
| *** mikecmpbll has quit IRC | 20:01 | |
| timburke | zaitcev: i blame pickle. my best guess is that when it wrote out some random bytes over the builder file, it happened to write something that pickle interpreted like True[3] or something | 20:12 |
| zaitcev | umm, okay | 20:12 |
| timburke | all the more reason for us to address https://bugs.launchpad.net/swift/+bug/1644387 | 20:13 |
| openstack | Launchpad bug 1644387 in OpenStack Object Storage (swift) "Ring builder files need a consistent format" [Undecided,New] | 20:13 |
| zaitcev | BTW, if you could review it, it would be great. I'm looking into account/container now and db.py used basically everywhere. | 20:14 |
| timburke | done! | 20:42 |
| *** e0ne has joined #openstack-swift | 20:44 | |
| *** e0ne has quit IRC | 20:49 | |
| zaitcev | thanks a lot! | 20:59 |
| zaitcev | You had a knack of finding much simpler solutions, so I expected something like "what is this pile of.." | 20:59 |
| zaitcev | waaaait a moment, I see a pile of comments | 21:00 |
| timburke | all nits. it's better to have it tested on py3 than not | 21:32 |
| timburke | i was close to just doing a +1, but then i figured out why we were doing byte strings in tests, and why it didn't matter | 21:33 |
| timburke | tdasilva: hmm... the storage1.error/storage1.log for the multinode tests... shouldn't that have more than just account-server? http://logs.openstack.org/05/581905/5/check/swift-multinode-rolling-upgrade/d3cab78/storage1.log | 21:34 |
| zaitcev | I did it because SQLite uses encoded (bytes) strings as arguments or returned values from selects and cursors. | 21:35 |
| zaitcev | Buuuut | 21:35 |
| zaitcev | or | 21:36 |
| zaitcev | well, I'm just afraid to become incompatible with the ondisk format, with just 1 small mis-step somewhere. | 21:36 |
| DHE | /dev/sdi 3.7T 210G 3.5T 6% /srv/node/slot3-phy9 // x30... good first lab machine... | 21:37 |
| timburke | zaitcev: it is indeed... :-( | 21:39 |
| timburke | don't worry about the follow-up; i'll take care of it. you go get some account/container servers going! | 21:40 |
| openstackgerrit | Tim Burke proposed openstack/swift master: DB follow-up https://review.openstack.org/608367 | 21:43 |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: Micro cleanup by using dict.clear https://review.openstack.org/608366 | 21:43 |
| timburke | bah! you beat me by a hair :P | 21:43 |
| *** chocolate-elvis has quit IRC | 21:49 | |
| *** itlinux has quit IRC | 21:50 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Move base64-decoding/unpickling into DatabaseBroker https://review.openstack.org/608369 | 22:00 |
| timburke | one more follow-up. zaitcev, we might want to be careful that we always have byte strings coming out of make_tuple_for_pickle... if only to ensure we can read py2 pending files on py3 | 22:02 |
| zaitcev | Hmm. | 22:04 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Move base64-decoding/unpickling into DatabaseBroker https://review.openstack.org/608369 | 22:08 |
| zaitcev | Indeed, a record supplied to put_record() is full of native strings and integers. | 22:10 |
| zaitcev | Making a tuple of it does nothing to cast. | 22:10 |
| timburke | yeah -- i meant the items *in* the tuple. native strings might be ok... but only if we can also handle bytes when we find 'em | 22:13 |
| timburke | or, we could say that py3 servers need to be *new* servers, similar to what rledisez and alecuyer are doing with losf -- the new thing will accept new data and can handle replication (and maybe even rsynced dbs), but it's not willing to read old pending files | 22:16 |
| timburke | but we'd need to be very clear about that | 22:16 |
| timburke | i've got this feeling like it wouldn't be *so* bad to be willing to read py2 pendings, though. just gotta make sure it's tested | 22:17 |
| zaitcev | Should we open a bug as a reminder? | 22:18 |
| *** itlinux has joined #openstack-swift | 22:19 | |
| *** itlinux has quit IRC | 22:22 | |
| zaitcev | Oooh great, now it's the upgrade gate. | 22:39 |
| zaitcev | So, basically recheck until it succeeds | 23:14 |
| notmyname | now you know how to openstack! ;-) | 23:22 |
| *** gyee has quit IRC | 23:31 | |
| openstackgerrit | Merged openstack/swift master: py3: Adapt db.py https://review.openstack.org/581905 | 23:55 |
| openstackgerrit | Merged openstack/swift master: DB follow-up https://review.openstack.org/608367 | 23:56 |
| openstackgerrit | Merged openstack/swift master: Move base64-decoding/unpickling into DatabaseBroker https://review.openstack.org/608369 | 23:56 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!