| *** aerwin has quit IRC | 00:06 | |
| *** annegentle has quit IRC | 00:18 | |
| openstackgerrit | Merged openstack/swift: Add extra_header_count to document and config. https://review.openstack.org/209525 | 00:25 |
|---|---|---|
| openstackgerrit | Merged openstack/swift: do container listing updates in another (green)thread https://review.openstack.org/189080 | 00:27 |
| openstackgerrit | Merged openstack/swift: Cleanup nits from container_update_timeout https://review.openstack.org/204813 | 00:27 |
| clayg | who look out - puts just got faster! | 00:30 |
| *** tsg has quit IRC | 00:33 | |
| *** bapalm has joined #openstack-swift | 00:37 | |
| notmyname | woohoo | 00:37 |
| *** bapalm has quit IRC | 00:39 | |
| *** minwoob has quit IRC | 00:40 | |
| *** gyee has quit IRC | 00:41 | |
| *** nakagawamsa has joined #openstack-swift | 00:55 | |
| *** robefran has joined #openstack-swift | 00:57 | |
| *** bill_az has quit IRC | 01:06 | |
| *** nakagawamsa has quit IRC | 01:10 | |
| *** chsc has quit IRC | 01:14 | |
| *** annegentle has joined #openstack-swift | 01:14 | |
| *** kota_ has quit IRC | 01:27 | |
| *** robefran has quit IRC | 01:30 | |
| *** logan2 has quit IRC | 02:08 | |
| *** logan2 has joined #openstack-swift | 02:09 | |
| *** darrenc is now known as darrenc_afk | 02:12 | |
| *** nakagawamsa has joined #openstack-swift | 02:21 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: pep8 fix: assertEquals -> assertEqual https://review.openstack.org/209649 | 02:24 |
| *** hezhiqiang has joined #openstack-swift | 02:24 | |
| *** km has quit IRC | 02:28 | |
| *** km has joined #openstack-swift | 02:28 | |
| *** tsg has joined #openstack-swift | 02:31 | |
| *** tsg has quit IRC | 02:36 | |
| *** aagrawal has quit IRC | 02:43 | |
| *** jrichli has joined #openstack-swift | 02:50 | |
| *** km has quit IRC | 03:03 | |
| *** km has joined #openstack-swift | 03:04 | |
| *** silor has joined #openstack-swift | 03:07 | |
| *** sanchitmalhotra has joined #openstack-swift | 03:07 | |
| *** logan2 has quit IRC | 03:14 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 03:24 | |
| *** dmorita has quit IRC | 03:25 | |
| *** dmorita has joined #openstack-swift | 03:25 | |
| *** sanchitmalhotra has quit IRC | 03:26 | |
| peluse | charz, you there? | 03:28 |
| *** sakaYK has joined #openstack-swift | 03:29 | |
| charz | peluse: yeah, | 03:29 |
| *** breitz has joined #openstack-swift | 03:29 | |
| peluse | ahh, just added a note to the bug :) | 03:30 |
| peluse | question actually, take a quick look | 03:30 |
| peluse | but how do you end up with just one token that works for both PACOs? You cmd line has just one right and don't you need to get one from each PACO? | 03:30 |
| *** logan2 has joined #openstack-swift | 03:31 | |
| charz | peluse: I think the memcached will take that part. proxies query memcached to see the token is expired or not. | 03:32 |
| peluse | so still don't quite understand, you have 2 totally independnt PACO nodes right? | 03:33 |
| peluse | basically 2 clusters | 03:33 |
| charz | peluse: no, that's one cluster with 2 PACO nodes. | 03:33 |
| peluse | ohhhh | 03:33 |
| charz | LOL | 03:34 |
| peluse | so the proxy service isn't being used on one of them? | 03:34 |
| peluse | if that's the case then how does ssbench talk to the other node? I'm so confused :) | 03:35 |
| *** marzif__ has joined #openstack-swift | 03:35 | |
| *** darrenc_afk is now known as darrenc | 03:36 | |
| charz | let me explain it more clear. here is a cluster with two PACOs node that include proxy/account/container/object servers are running on it. | 03:36 |
| peluse | ok, go on... | 03:36 |
| charz | And I setup two benchmark nodes(other VMs), ssbench installed on both node. | 03:37 |
| peluse | whoa | 03:37 |
| charz | I get a token from one proxy of PACOs node, and then trigger ssbench-master to against two proxies(PACOs). | 03:39 |
| charz | I skip the load balancer in front of cluster. | 03:39 |
| peluse | yeah, I get that there's no LB | 03:40 |
| charz | And here is one ssbench-master in each benchmark node. | 03:41 |
| *** dmorita has quit IRC | 03:42 | |
| charz | I uses the command I post in gist. | 03:42 |
| peluse | I think the thing I'm missing here is that I've not manually setup a cluster w/>1 proxy on it before. I either just use SAIO or the SwiftStack product. Guess I need to RTFM | 03:43 |
| peluse | somy current setup is just 2 single node clusters which is why I'm confused by having just one token | 03:43 |
| peluse | gotta run, will get my stuff fixed up sometime between now and the hackathon I hope. Have a busy 2 days before heading out there... | 03:45 |
| peluse | thanks! | 03:45 |
| charz | peluse: yeah, I'll confuse that too. | 03:45 |
| mattoliverau | When I used ssync I had to set up keystone so all the proxies could use the same token | 03:45 |
| charz | peluse: np, just ping me if you have problems. | 03:46 |
| *** aagrawal has joined #openstack-swift | 03:50 | |
| *** ppai has joined #openstack-swift | 04:03 | |
| mattoliverau | *ssbench (lol too much swift fingers autocomplete) | 04:03 |
| *** zaitcev has quit IRC | 04:10 | |
| *** early has quit IRC | 04:16 | |
| *** early has joined #openstack-swift | 04:18 | |
| *** sakaYK has quit IRC | 04:19 | |
| *** sakaYK has joined #openstack-swift | 04:20 | |
| *** dmorita has joined #openstack-swift | 04:24 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Restrict account_autocreate behavior with allow_account_management https://review.openstack.org/209750 | 04:25 |
| *** tsg has joined #openstack-swift | 04:25 | |
| *** sakaYK has quit IRC | 04:25 | |
| *** sakaYK has joined #openstack-swift | 04:26 | |
| *** breitz has quit IRC | 04:35 | |
| ho | notmyname: thanks for the explanation. I created a patch for it. :-) | 04:40 |
| notmyname | ho: interesting. thanks | 04:40 |
| *** annegentle has quit IRC | 04:41 | |
| notmyname | ho: I say "interesting" because I'm not sure that it's what is expected | 04:41 |
| notmyname | i'll take a look | 04:41 |
| *** jrichli has quit IRC | 04:41 | |
| notmyname | (later) | 04:41 |
| *** tsg has quit IRC | 04:43 | |
| *** tsg has joined #openstack-swift | 04:44 | |
| *** ppai has quit IRC | 04:50 | |
| charz | ho: the closes-bug link is broken or that's incorrect bug id(1482041)? | 04:52 |
| ho | charz: it looks correct. https://bugs.launchpad.net/swift/+bug/1482041 | 05:02 |
| openstack | ho: Error: malone bug 1482041 not found | 05:02 |
| *** ppai has joined #openstack-swift | 05:04 | |
| ho | notmyname: thanks! | 05:09 |
| *** itlinux has quit IRC | 05:09 | |
| StevenK | charz: That bug is most likely private | 05:10 |
| *** hrou has quit IRC | 05:10 | |
| StevenK | Which means the bot can't do anything to it | 05:10 |
| charz | StevenK: yeah, ho just told me. That's private bug. | 05:10 |
| charz | StevenK: thx | 05:11 |
| *** SkyRocknRoll_ has quit IRC | 05:14 | |
| *** sakaYK has quit IRC | 05:14 | |
| *** sanchitmalhotra has joined #openstack-swift | 05:20 | |
| *** sanchitmalhotra1 has quit IRC | 05:22 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 05:29 | |
| *** sanchitmalhotra has quit IRC | 05:31 | |
| *** ppai has quit IRC | 05:33 | |
| *** ppai has joined #openstack-swift | 05:45 | |
| *** ho has quit IRC | 05:46 | |
| *** ho has joined #openstack-swift | 05:46 | |
| openstackgerrit | Merged openstack/swift: Fix TypeError if backend response doesn't have expected headers https://review.openstack.org/209370 | 05:51 |
| openstackgerrit | Merged openstack/swift: Having said H, I, J, we ought to say K https://review.openstack.org/209677 | 05:56 |
| openstackgerrit | Merged openstack/swift: pep8 fix: assertEquals -> assertEqual https://review.openstack.org/209642 | 05:56 |
| openstackgerrit | Merged openstack/swift: pep8 fix: assertEquals -> assertEqual https://review.openstack.org/209643 | 05:57 |
| *** mahatic has joined #openstack-swift | 05:57 | |
| *** sakaYK has joined #openstack-swift | 06:13 | |
| *** sanchitmalhotra has joined #openstack-swift | 06:14 | |
| *** sanchitmalhotra1 has quit IRC | 06:16 | |
| *** sakaYK has quit IRC | 06:19 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:20 | |
| *** sakaYK has joined #openstack-swift | 06:21 | |
| *** dmorita has quit IRC | 06:22 | |
| *** sakaYK has quit IRC | 06:25 | |
| *** sakaYK has joined #openstack-swift | 06:26 | |
| *** marzif__ has quit IRC | 06:39 | |
| *** sakaYK has quit IRC | 06:44 | |
| *** tsg has quit IRC | 06:45 | |
| *** sakaYK has joined #openstack-swift | 06:45 | |
| *** sakaYK has quit IRC | 06:50 | |
| *** sakaYK has joined #openstack-swift | 06:51 | |
| *** sakaYK has quit IRC | 06:57 | |
| *** sakaYK has joined #openstack-swift | 06:58 | |
| *** sakaYK has quit IRC | 07:03 | |
| *** sakaYK has joined #openstack-swift | 07:04 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift-specs: Update sharding spec, notably CountingTrie and shrinking https://review.openstack.org/201000 | 07:06 |
| mattoliverau | ^ latest SPEC update (or braindump) now including initial version of shrinking sharded containers (collapsing). | 07:07 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Make swift-ring-builder filename usage more consistent https://review.openstack.org/209787 | 07:16 |
| *** sanchitmalhotra1 has joined #openstack-swift | 07:21 | |
| *** sanchitmalhotra has quit IRC | 07:23 | |
| *** sakaYK has quit IRC | 07:23 | |
| *** sakaYK has joined #openstack-swift | 07:24 | |
| *** sakaYK has quit IRC | 07:30 | |
| *** sakaYK has joined #openstack-swift | 07:30 | |
| *** proteusguy has quit IRC | 07:34 | |
| *** sakaYK has quit IRC | 07:36 | |
| *** sanchitmalhotra has joined #openstack-swift | 07:36 | |
| *** sakaYK has joined #openstack-swift | 07:37 | |
| *** sanchitmalhotra1 has quit IRC | 07:38 | |
| openstackgerrit | Jamie Lennox proposed openstack/python-swiftclient: Add some logging debug for authentication https://review.openstack.org/209800 | 07:42 |
| *** sakaYK has quit IRC | 07:43 | |
| *** sakaYK_ has joined #openstack-swift | 07:43 | |
| *** sakaYK_ has quit IRC | 07:48 | |
| *** sakaYK has joined #openstack-swift | 07:49 | |
| openstackgerrit | Jamie Lennox proposed openstack/python-swiftclient: Add some logging debug for authentication https://review.openstack.org/209800 | 07:50 |
| *** proteusguy has joined #openstack-swift | 07:52 | |
| *** sakaYK has quit IRC | 07:53 | |
| *** sakaYK has joined #openstack-swift | 07:54 | |
| *** geaaru has joined #openstack-swift | 07:54 | |
| *** sakaYK has quit IRC | 07:56 | |
| *** jordanP has joined #openstack-swift | 07:57 | |
| *** sakaYK has joined #openstack-swift | 07:57 | |
| *** sakaYK has quit IRC | 08:00 | |
| *** sakaYK has joined #openstack-swift | 08:00 | |
| *** sakaYK has quit IRC | 08:05 | |
| *** sakaYK has joined #openstack-swift | 08:05 | |
| *** ttrumm has joined #openstack-swift | 08:09 | |
| *** saltsa_ has quit IRC | 08:13 | |
| *** jistr has joined #openstack-swift | 08:14 | |
| openstackgerrit | kenichiro matsuda proposed openstack/swift: Fix shebang of commands https://review.openstack.org/209811 | 08:15 |
| *** SkyRocknRoll has quit IRC | 08:39 | |
| *** SkyRocknRoll has joined #openstack-swift | 08:39 | |
| *** marzif__ has joined #openstack-swift | 08:54 | |
| *** joeljwright has joined #openstack-swift | 08:54 | |
| *** ChanServ sets mode: +v joeljwright | 08:54 | |
| *** mahatic has quit IRC | 08:57 | |
| *** SkyRocknRoll has quit IRC | 08:59 | |
| *** mahatic has joined #openstack-swift | 08:59 | |
| *** SkyRocknRoll has joined #openstack-swift | 08:59 | |
| mahatic | acoles_away: morning! | 09:00 |
| acoles_away | mahatic: hi | 09:01 |
| * acoles_away notices that his away-nick bot is broken | 09:01 | |
| mahatic | ah! was wondering if this is a good time to ask. thanks for that info! | 09:01 |
| *** acoles_away is now known as acoles | 09:02 | |
| acoles | mahatic: ok i have adopted my alter-ego | 09:02 |
| mahatic | acoles: :) I'm trying to figure out a way to move forward on those probe tests. For what jrichli asked on etherpad "Is there any use case (aside from testing) that we would need to decrypt what is received from direct_get_object? Also, is it ok to skip those tests altogether when using encryption" | 09:04 |
| mahatic | acoles: from my knowledge, I don't see a need to test on direct_get_object since it's for swift backend | 09:04 |
| *** sakaYK has quit IRC | 09:05 | |
| mahatic | acoles: to test encryption on* If you think otherwise, please let me know and maybe I could go in that direction | 09:06 |
| *** sanchitmalhotra1 has joined #openstack-swift | 09:08 | |
| mahatic | maybe it is needed somewhere, but I'm not familiar with such a usecase | 09:10 |
| *** sanchitmalhotra has quit IRC | 09:10 | |
| acoles | mahatic: hmmm, i have thought a bit about that and still not sure how best to proceed. probe tests assume a certain configuration (basically, the SAIO setup). Question is, does encyption become part of that 'standard' setup? | 09:12 |
| *** sakaYK has joined #openstack-swift | 09:13 | |
| acoles | mahatic: so far my thinking is 'no', at least not encryption enabled by default. I could be wrong, but my guess we will not move to SAIO encrypting by default. | 09:14 |
| acoles | mahatic: some probe tests are conditional e.g. erasure code specific tests which Skip if no EC policy is found. | 09:14 |
| mahatic | acoles: oh, that's interesting - conditional probe tests | 09:15 |
| acoles | mahatic: but i'm not a fan of making all existing probe tests conditional on encryption NOT being enabled. So I still think that having a way to dynamically enable encryption for specific (new perhaps) probe tests is a better way forwards. | 09:16 |
| mahatic | acoles: that sounds right | 09:17 |
| acoles | i.e. for testing, crypto middleware could be in the proxy pipeline but not enabled by default. that way all probe existing tests should pass. | 09:17 |
| acoles | then we have a mechanism for a test to enable crypto eg by setting a header that our trivial keymaster would detect. | 09:18 |
| mahatic | right | 09:18 |
| acoles | But then I am wondering how that sits w.r.t. functional testing - does that mean that functional tests have to set that header? yuk :/ | 09:19 |
| mahatic | why not? coz there will be too many? | 09:20 |
| *** haomaiwang has quit IRC | 09:21 | |
| acoles | mahatic: just need take a call. bbiab | 09:21 |
| *** haomaiwang has joined #openstack-swift | 09:21 | |
| * mahatic looksup bbiab - got it | 09:22 | |
| * mahatic is only used to brb ;) | 09:22 | |
| acoles | mahatic: sorry, back now | 09:29 |
| mahatic | acoles: okay, np | 09:29 |
| acoles | mahatic: We do now have a way for functional tests to target a specific policy, so if we made out triv keymaster only enable crypto for a specific policy then that might be a way. | 09:29 |
| acoles | then crypto specific probe tests use that policy or skip if its not found. | 09:30 |
| mahatic | acoles: okay. we will still need to tests specific to encryption, in a case where encryption is enabled at the account level irrespective of the policy I think? | 09:31 |
| mahatic | need tests* | 09:31 |
| mahatic | s/encryption/crypto | 09:33 |
| mahatic | acoles: I mean how do we test those? | 09:35 |
| acoles | mahatic: not sure i follow. probe or functional test? | 09:35 |
| mahatic | acoles: crypto could be enabled not just based on policy, correct? | 09:36 |
| mahatic | "Vinz also checks the account metadata for a metadata item350 | 09:38 |
| mahatic | 'X-Account-Sysmeta-Vinz-Encrypt: always' that a sys admin may have set. If351 | 09:38 |
| mahatic | present Vinz will specify object encryption regardless of the container352 | 09:38 |
| mahatic | policy." | 09:38 |
| mahatic | acoles: excuse those line numbers | 09:38 |
| mahatic | acoles: do I make sense? | 09:39 |
| acoles | oic. yes. so let me clarify - func tests should pass against *any* cluster. some func tests skip if the cluster doesnt have certain features. So yes, whatever keymaster your cluster uses, func tests should pass. | 09:40 |
| mahatic | okay | 09:41 |
| acoles | What i was trying to address is when you have an SAIO with our triv keymaster that we dont want enabled *all the time* (cos it would break probe tests) - then how do we force func tests to exercise crypto? | 09:41 |
| acoles | And I was thinking we could do that by running the func tests using a specific policy, and the policy would trigger the trivial keymaster to enable crypto. | 09:42 |
| mahatic | right, I think I moved a lil away from it | 09:42 |
| mahatic | correct | 09:42 |
| acoles | So I could e.g. func test my SAIO without crypto and also with crypto enabled by choosing the func test policy. | 09:43 |
| mahatic | acoles: yes, basically it does (should) cover all of that I mentioned. thanks for the explanation! | 09:44 |
| acoles | mahatic: anyway, in terms of existing probe tests, i dont think we should rush to change them to work with crypto enabled. If we write new crypto specific probe tests we should make them skip if crypto middleware is not installed. | 09:45 |
| mahatic | acoles: okay. sounds right | 09:46 |
| acoles | mahatic: i still advocate adding ciphertext checking to the unit tests in the near term | 09:46 |
| acoles | as we discussed before, that would be great. | 09:47 |
| mahatic | acoles: ah. so in the same unit test that I refered? or as you suggested later, a new test? | 09:47 |
| *** hezhiqia_ has joined #openstack-swift | 09:47 | |
| *** hezhiqi__ has joined #openstack-swift | 09:48 | |
| acoles | mahatic: i think jrichli said the existing test mocked the crypto engine, so yeah maybe a new unit test class that would be similar but use a real instance of the crypto engine. | 09:49 |
| *** hezhiqiang has quit IRC | 09:49 | |
| acoles | mahatic: or look at why the cryto engine needs to be mocked in existing? is that necessary? (I need to go study that some more) | 09:49 |
| openstackgerrit | Hiroshi Miura proposed openstack/python-swiftclient: change deprecated assertEquals to assertEqual https://review.openstack.org/208331 | 09:50 |
| acoles | mahatic: i'll add some notes to ether pad based on what we just discussed. | 09:50 |
| mahatic | acoles: sure, I haven't looked into the why part yet, will do that | 09:50 |
| mahatic | acoles: sure, that'd be great | 09:51 |
| acoles | cool | 09:51 |
| acoles | mahatic: are you coming to austin next week? | 09:51 |
| mahatic | acoles: nope! :( but for any discussion around this, I'd love jump in through skype if that's possible | 09:52 |
| *** hezhiqia_ has quit IRC | 09:52 | |
| acoles | mahatic: ah, shame. yes we should try to make that happen. | 09:53 |
| *** silor has quit IRC | 09:54 | |
| *** sanchitmalhotra has joined #openstack-swift | 09:58 | |
| *** sanchitmalhotra1 has quit IRC | 10:01 | |
| *** jamespage has joined #openstack-swift | 10:01 | |
| *** joeljwright has quit IRC | 10:04 | |
| *** hezhiqiang has joined #openstack-swift | 10:05 | |
| *** hezhiqi__ has quit IRC | 10:06 | |
| *** hezhiqia_ has joined #openstack-swift | 10:07 | |
| *** hezhiqiang has quit IRC | 10:10 | |
| *** kei_yama has quit IRC | 10:15 | |
| *** joeljwright has joined #openstack-swift | 10:16 | |
| *** ChanServ sets mode: +v joeljwright | 10:16 | |
| *** hezhiqia_ has quit IRC | 10:19 | |
| *** sakaYK_ has joined #openstack-swift | 10:21 | |
| *** sakaYK has quit IRC | 10:22 | |
| *** ho has quit IRC | 10:23 | |
| *** iepupp_ has joined #openstack-swift | 10:24 | |
| iepupp_ | why https://review.openstack.org/#/q/topic:bp/object-patch-support,n,z was Abandoned? | 10:26 |
| *** eandersson has joined #openstack-swift | 10:26 | |
| *** sakaYK_ has quit IRC | 10:42 | |
| *** SkyRocknRoll has quit IRC | 11:08 | |
| mattoliverau | iepupp: first one because back in 2013, if there was a negative score and no action for a few weeks it gets auto abandoned, the other 2 were abandoned by the author (a long time ago) | 11:13 |
| *** SkyRocknRoll has joined #openstack-swift | 11:20 | |
| *** km has quit IRC | 11:29 | |
| *** km has joined #openstack-swift | 11:29 | |
| *** eandersson has quit IRC | 11:35 | |
| *** eandersson has joined #openstack-swift | 11:36 | |
| openstackgerrit | Jamie Lennox proposed openstack/python-swiftclient: Add some logging debug for authentication https://review.openstack.org/209800 | 11:36 |
| *** iepupp_ has quit IRC | 11:41 | |
| *** silor has joined #openstack-swift | 11:45 | |
| zigo | https://buildd.debian.org/status/fetch.php?pkg=python-pyeclib&arch=arm64&ver=1.0.8-1&stamp=1438770260 | 11:50 |
| zigo | any idea what's going on? | 11:50 |
| *** mahatic_ has joined #openstack-swift | 11:57 | |
| *** mahatic has quit IRC | 11:57 | |
| *** mahatic_ has quit IRC | 11:59 | |
| *** mahatic_ has joined #openstack-swift | 12:00 | |
| *** bill_az has joined #openstack-swift | 12:07 | |
| jordanP | zigo, you are packaging python-pyeclib-1.0.8 but you have a system with liberasurecode-dev (1.0.5-2) | 12:17 |
| jordanP | zigo, this looks wrong to me | 12:18 |
| *** yuanzz has quit IRC | 12:34 | |
| *** bapalm has joined #openstack-swift | 12:44 | |
| *** zul has joined #openstack-swift | 12:46 | |
| zigo | jordanP: Makes sense! Good catch. I'll fix the build-deps. | 12:49 |
| *** marzif__ has quit IRC | 12:53 | |
| *** marzif__ has joined #openstack-swift | 12:54 | |
| *** nakagawamsa has quit IRC | 12:55 | |
| *** km has quit IRC | 12:56 | |
| tdasilva | just ran into a similar issue with liberasurecode-dev: http://ur1.ca/ndblr | 12:58 |
| tdasilva | clayg, peluse ^^^ | 12:58 |
| *** aagrawal has quit IRC | 13:02 | |
| *** tsg has joined #openstack-swift | 13:05 | |
| portante | tdasilva, notmyname: check out https://github.com/elastic/packetbeat, for those that want to integrate with ElasticSearch, might be cool for swift | 13:06 |
| *** mahatic_ has quit IRC | 13:12 | |
| *** robefran has joined #openstack-swift | 13:15 | |
| *** ppai has quit IRC | 13:19 | |
| *** hrou has joined #openstack-swift | 13:21 | |
| *** dustins has joined #openstack-swift | 13:22 | |
| *** thumpba has joined #openstack-swift | 13:23 | |
| *** thumpba_ has joined #openstack-swift | 13:26 | |
| *** vinsh_ has joined #openstack-swift | 13:27 | |
| *** doxavore has joined #openstack-swift | 13:29 | |
| *** vinsh has quit IRC | 13:29 | |
| doxavore | is there any documentation on what causes quarantined objects (as reported by swift-recon)? | 13:30 |
| *** thumpba has quit IRC | 13:30 | |
| *** marcusvrn has joined #openstack-swift | 13:30 | |
| *** thumpba_ has quit IRC | 13:30 | |
| *** petertr7_away is now known as petertr7 | 13:31 | |
| *** thumpba has joined #openstack-swift | 13:36 | |
| *** rledisez has joined #openstack-swift | 13:36 | |
| *** proteusguy has quit IRC | 13:37 | |
| openstackgerrit | Eran Rom proposed openstack/swift: Container-Sync to iterate only over synced containers https://review.openstack.org/205803 | 13:37 |
| *** vinsh has joined #openstack-swift | 13:39 | |
| rledisez | hi onovy | 13:39 |
| rledisez | onovy: i have a review ongoing (https://review.openstack.org/#/c/192186/) and clayg added a comment about yours | 13:39 |
| rledisez | onovy: basically, he's suggesting to add the rsync module name in the ring, like you did for the port | 13:40 |
| rledisez | onovy: i was wondering what is your use case? because it could match mine too :) | 13:41 |
| *** vinsh_ has quit IRC | 13:42 | |
| *** SkyRocknRoll has quit IRC | 13:45 | |
| *** breitz has joined #openstack-swift | 13:49 | |
| *** proteusguy has joined #openstack-swift | 13:53 | |
| *** thurloat_isgone is now known as thurloat | 13:53 | |
| tsubic | are there any open source application examples of swift usage? I would like to just take a look at a programming model for using swift. Thanks | 13:57 |
| *** breitz has quit IRC | 13:58 | |
| *** dustins has quit IRC | 13:59 | |
| jamespage | jordanP, matching the liberasurecode versions appears to improve the unit testing situation in Debian - thanks for that pointer | 14:00 |
| jordanP | jamespage, that's good to know | 14:00 |
| *** mahatic has joined #openstack-swift | 14:01 | |
| jamespage | jordanP, I think we're seeing some endian type issues on two archs - just waiting to confirm that | 14:01 |
| jamespage | well at least tests fail on two big-endian architectures | 14:01 |
| jamespage | https://buildd.debian.org/status/package.php?p=python-pyeclib&suite=sid | 14:02 |
| jamespage | for reference | 14:02 |
| *** jrichli has joined #openstack-swift | 14:05 | |
| *** tsg has quit IRC | 14:10 | |
| *** joeljwright has quit IRC | 14:11 | |
| *** jlhinson has joined #openstack-swift | 14:14 | |
| *** dustins has joined #openstack-swift | 14:17 | |
| *** ttrumm has quit IRC | 14:21 | |
| *** saltsa has joined #openstack-swift | 14:22 | |
| *** joeljwright has joined #openstack-swift | 14:25 | |
| *** ChanServ sets mode: +v joeljwright | 14:25 | |
| openstackgerrit | Bill Huber proposed openstack/swift: pep8 fix: assertEquals -> assertEqual https://review.openstack.org/209652 | 14:30 |
| *** dmorita has joined #openstack-swift | 14:34 | |
| *** jasondot_ is now known as jasondotstar | 14:38 | |
| *** breitz has joined #openstack-swift | 14:41 | |
| *** breitz has joined #openstack-swift | 14:41 | |
| *** marzif__ has quit IRC | 14:43 | |
| *** marzif__ has joined #openstack-swift | 14:44 | |
| *** dustins_ has joined #openstack-swift | 14:45 | |
| *** annegentle has joined #openstack-swift | 14:45 | |
| *** MVenesio has joined #openstack-swift | 14:45 | |
| *** dustins has quit IRC | 14:47 | |
| MVenesio | Hi guys, i'm running swift "Juno" version in 4 Ubuntu Boxes with 8 disks each. And i'm seeing [xfsaild(sdx] cousing 90% of I/O in all disks, while swift process are generating just 20% of I/O | 14:48 |
| *** delattec has joined #openstack-swift | 14:48 | |
| MVenesio | Any idea about this ? | 14:48 |
| *** bill_az has quit IRC | 14:54 | |
| notmyname | good morning | 14:56 |
| *** bill_az has joined #openstack-swift | 14:57 | |
| *** Fin1te has joined #openstack-swift | 14:58 | |
| MVenesio | good morning | 14:59 |
| *** bapalm_ has joined #openstack-swift | 15:00 | |
| *** dustins_ is now known as dustins | 15:00 | |
| *** bapalm has quit IRC | 15:01 | |
| openstackgerrit | Bill Huber proposed openstack/swift: pep8 fix: assertEquals -> assertEqual https://review.openstack.org/209976 | 15:02 |
| *** bill_az has quit IRC | 15:04 | |
| *** lpabon has joined #openstack-swift | 15:04 | |
| *** bill_az has joined #openstack-swift | 15:05 | |
| notmyname | MVenesio: I don't knwo about that. quick google searches don't reveal much. perhaps ahale or donagh might have seen something like that and have ideas | 15:05 |
| MVenesio | notmyname: perfect i'll ask them, thanks | 15:06 |
| *** zul has quit IRC | 15:10 | |
| *** gyee has joined #openstack-swift | 15:14 | |
| *** geaaru has quit IRC | 15:21 | |
| *** breitz has quit IRC | 15:22 | |
| *** breitz has joined #openstack-swift | 15:22 | |
| *** lcurtis has joined #openstack-swift | 15:30 | |
| *** minwoob has joined #openstack-swift | 15:31 | |
| *** blmartin has joined #openstack-swift | 15:33 | |
| mahatic | acoles: re do we need to mock crypto in enginer test_encrypter: It makes a call to Encrypter (class) and mocks crypto engine, so there isn't a way the actual encryption/decryption is tested | 15:33 |
| *** janonymous_ has joined #openstack-swift | 15:34 | |
| mahatic | jrichli: why exactly should the ecrypter be ignorant of crypto engine? | 15:34 |
| mahatic | in test_encrypter | 15:34 |
| mahatic | acoles: jrichli: I don't particularly see a need to mock it, when we might call the real crypto and do real testing of encryption/decryption? | 15:35 |
| *** bill_az has quit IRC | 15:35 | |
| *** bill_az has joined #openstack-swift | 15:36 | |
| *** nadeem has joined #openstack-swift | 15:39 | |
| *** tongli has joined #openstack-swift | 15:44 | |
| *** lpabon has quit IRC | 15:44 | |
| *** silor has quit IRC | 15:47 | |
| lcurtis | hello all...i swapped out a disk yesterday and am seeing ERROR Container update failed (saving for async update later): 404 response from x.x.x.x:6001/disk-x | 15:49 |
| *** sanchitmalhotra1 has joined #openstack-swift | 15:50 | |
| lcurtis | container server is quite busy iowait of 2-3....would that be overall reason? | 15:51 |
| jrichli | mahatic: we do want to add a test that integrates with the real crypto. but we want to do this in such a way that a different crypto algorithm could be used later, if desired, with no modifications to the encrypter test. | 15:51 |
| jrichli | mahatic: but we can easily do this. | 15:51 |
| *** sanchitmalhotra has quit IRC | 15:52 | |
| *** petertr7 is now known as petertr7_away | 15:52 | |
| mahatic | jrichli: but what exactly are testing with this test_encrypter that won't be tested when crypto engine is not mocked? Even if there was a diff crypto algo used later, we'll have new tests per requirement right? | 15:53 |
| mahatic | are we* | 15:54 |
| *** janonymous_ has quit IRC | 15:54 | |
| *** nadeem has quit IRC | 15:55 | |
| *** SkyRocknRoll has joined #openstack-swift | 15:57 | |
| jrichli | mahatic: I don't think we don't need a new test_encrypter class. Just add a new test case in the current file that integrates with the real crypto. You might have to change how the mock is specified, since it is globally applied currently. | 15:58 |
| jrichli | mahatic: yes, somebody could always modify the encrypter tests when crypto changes, but that isn't making crypto very modular. | 15:58 |
| openstackgerrit | Romain LE DISEZ proposed openstack/swift: Allows the replicator to rsync to a device-based module https://review.openstack.org/192186 | 15:59 |
| *** rledisez has quit IRC | 15:59 | |
| jrichli | mahatic: it is a best practice to keep items such as the crypto alg (that you predict might change later) modularized | 15:59 |
| jrichli | mahatic: I dont think it will be difficult to keep it modular | 16:01 |
| mahatic | jrichli: um that's true. my only question is, what are we specifically testing by mocking crypto engine? If it is only making sure that a call to Encrypter goes fine, can we not integrate it with real crypto itself? | 16:01 |
| jrichli | mahatic: the encrypter is a pretty large class with logic about how the headers are to be encrypted. There is a lot that can be unit tested in just that class alone. | 16:03 |
| *** marzif__ has quit IRC | 16:03 | |
| jrichli | mahatic: currently, all the tests there are asserting behavior in just the encrypter class. trust me, they can fail if the encyrpter is not donig its job - has happened to me for sure! | 16:03 |
| mahatic | jrichli: yes I did go through the whole thing, although it does include the logic about how the headers are to be encrypter, we're still faking the encryption part | 16:04 |
| mahatic | jrichli: ah alright | 16:04 |
| acoles | jrichli: but the current crypto engine can't ever go away, correct? so how does it hurt to incorporate it in unit tests? | 16:04 |
| jrichli | mahatic: it does not hurt. we are suggesting that a test case be added to incorporate it. | 16:05 |
| mahatic | acoles: exactly | 16:05 |
| mahatic | jrichli: do you mean remove the mock part and incorporate the real crypto? | 16:06 |
| acoles | jrichli: it was me that typed that :P | 16:06 |
| mahatic | :D | 16:06 |
| jrichli | acoles: yes, i just noticed that. :-) | 16:06 |
| acoles | and yes that is what i meant | 16:06 |
| acoles | then if another engine comes along it may get some unit tests of its own but the existing tests have to still work | 16:07 |
| jrichli | mahatic, acoles: i am a bit of a unit test purist. I like having tests that test one thing - like encrypter. I think the integrated test case is a good idea to have in addition | 16:07 |
| hrou | acoles, mahatic, I think the idea is indeed to incorporate it while keeping the encrypter unit tests independent, best of both worlds ? : ) | 16:07 |
| jrichli | the crypto alg has unit tests of its own | 16:08 |
| jrichli | we did talk about *someday* somebody *might* want to have the crypto alg be configurable. Not to bring up a bid debate, | 16:08 |
| mahatic | jrichli: acoles: in addition? meaning we're gonna have to repeat all the unit tests with real crypto, correct? | 16:08 |
| jrichli | but it would be good to not make decisions that will lead to that being impossible | 16:09 |
| *** potato_farmer has joined #openstack-swift | 16:09 | |
| acoles | jrichli: hrou: maybe. i like testing the "whole" as much as possible. its too easy to build modules that are happy with their own tests but diverge from working together. | 16:09 |
| potato_farmer | Hi. I am trying to remove a swift storage node. I have balance, partitions, and weight down to 0. I have waited several hours between each try. But anytime I try to remove the IP, I get: "No partitions could be reassigned. Either none need to be or none can be due to min_part_hours [1]." Has anyone run into this before? | 16:10 |
| jrichli | mahatic: no, I figured just a basic PUT test case should be enough to test the integration. | 16:10 |
| mahatic | jrichli: oh I see | 16:10 |
| hrou | acoles, true, it really depends on the attitude you take to unit tests, sometimes folks strive to keep them as independent as possible. They're fairly isolated here though ! Probably can just start by adding a direct crypto algo test (and I agree this is already fairly module, if for example in the future we support 3DES in the swift engine) | 16:11 |
| lcurtis | potato_farmer what is your min_part_hours set to | 16:11 |
| *** jistr has quit IRC | 16:12 | |
| potato_farmer | lcurtis: I am still pretty new to this, how can I check? | 16:14 |
| potato_farmer | lcurtis: based on the error, isn't it [1] ? | 16:16 |
| jrichli | mahatic acoles: I am open to changing the tests to all be more integrated, if you think that is the best direction | 16:17 |
| acoles | hrou: i just see existing tests where we could easily verify that stuff got encrypted/decrypted correctly, except the engine is mocked out. So instead we'll have to write another set of test that would either duplicate or not have complete coverage. | 16:17 |
| acoles | jrichli: happy to debate next week. i know folks have different views on test strategies. | 16:18 |
| *** jordanP has quit IRC | 16:18 | |
| mahatic | acoles: jrichli I won't be around next week! :\ debate here next week, on IRC :P | 16:19 |
| *** doxavore has quit IRC | 16:19 | |
| jrichli | acoles mahatic: no debate necessary on my end. if you all think that it's best to have all the test cases be integrated, I am on board. | 16:19 |
| jrichli | and that way, mahatic can start making the mods | 16:20 |
| mahatic | that's true though | 16:21 |
| hrou | acoles, back in my large software dev days (DB2) we were beaten with sticks when unit tests were not 100% isolated even if the end result was 100% more unit tests ; ) But its probably easy to change this all going forward either way ! | 16:21 |
| *** nadeem has joined #openstack-swift | 16:22 | |
| hrou | mahatic, speaking of encryption, quick question for you - did you have to play around with extra_header_count to achieve the test results cited on the etherpad ? | 16:22 |
| acoles | jrichli: mahatic : so a first step would be to remove the crypto mock and see what breaks in existing tests? | 16:22 |
| acoles | hrou: yep, understand, its just what you perceive to be a unit and i'm not convinced of the separation of the encrypter and crypto engine. The current engine isn't going to go away even if others come along, so I see it as the built in default if you like. | 16:24 |
| openstackgerrit | Bill Huber proposed openstack/swift: Replace python print operator with print function (pep H233, py33) https://review.openstack.org/210014 | 16:24 |
| lcurtis | potato_farmer run the swift-ring-builder command against your ring file: swift-ring-builder /etc/swift/object.builder | 16:25 |
| lcurtis | u should see: The minimum number of hours before a partition can be reassigned is X | 16:25 |
| potato_farmer | lcurtis: It shows "1" for account, container, and object | 16:26 |
| *** hezhiqiang has joined #openstack-swift | 16:27 | |
| lcurtis | potato_farmer so, something is odd then...hopefully one of the gurus that hang out here can assist | 16:27 |
| mahatic | acoles: I could do that | 16:27 |
| mahatic | jrichli: ^ | 16:28 |
| jrichli | mahatic: sounds good | 16:29 |
| *** chsc has joined #openstack-swift | 16:30 | |
| potato_farmer | lcurtis: I agree. I have removed nodes in the past without issues. | 16:30 |
| mahatic | hrou: oh I have had multiple issues, majorly pip, for whatever reason (I couldn't figure out all of that). But just downloading the code and running the tests should give you those results, did it not happen that way? | 16:31 |
| hrou | acoles, yep I agree the current engine will likely encapsulate the other algorithms (as it hard codes AES per say today) and maybe an account / policy setting dictates what to use. Either approach will work, I guess someone can argue if you had independent crypto algo unit tests you'd only have to run this X times (where X is the different algo's supported), as opposed to everything. Its a different story if the crypto was full | 16:31 |
| hrou | fledge middleware outside of swift. Maybe the answer is just do both : ) integrate and separate. But minor either way. | 16:31 |
| hrou | mahatic, thanks ! Yep so I actually had to set extra_header_count to some large number, question - did you run the tests fairly recently, I think only as of the last week or so (maybe 2) the account metadata encryption was sent for review ? | 16:33 |
| mahatic | hrou: a week and a half ago I believe. Sorry I can't correlate with the review you mentioned, will need to look | 16:34 |
| hrou | mahatic, there was a func test that set a bunch of metadata account headers => we add a bunch of headers in the keymaster, and my theory that was the cause. | 16:34 |
| *** hezhiqiang has quit IRC | 16:34 | |
| hrou | mahatic, oh no worries at all, let me dig into it more, maybe its just me ; ) Thanks a bunch though ! | 16:34 |
| acoles | hrou: interesting. maybe we bloat the backend headers enough to exceed the max header count. thanks for mentioning that! | 16:35 |
| jrichli | hrou: I know that I have never set this. But maybe setting it would fix something that has always failed for me, I don't know. | 16:35 |
| jrichli | hrou: you said that your failures match those that I have posted - that was even after you changed this config? | 16:36 |
| mahatic | hrou: I could try and reproduce this in another vm (tomm sometime) | 16:36 |
| hrou | jrichli, acoles - yep exactly, that was my thought, and on paper it definitely seems possible - by default we only allow 122 headers (this was recent change to the HTTP python lib) | 16:36 |
| hrou | acoles, if you recall clayg (I think, sorry if I'm wrong !) delivered a patch to main that allowed you to increase this through swift.conf, using extra_header_count and that did the trick for me. | 16:37 |
| jrichli | hrou: what test failed before that now succeeds? or do we have a test that is asserting this behavior? | 16:38 |
| hrou | mahatic, sure that'd be great, make sure to pull down all the recent reviews and rebase to latest (I did it manually) there was only a minor conflict, let me know if you have any issues. | 16:38 |
| *** annegentle has quit IRC | 16:38 | |
| jrichli | hrou mahatic: I have not yet rebased, so maybe it was from master that somebody added a test to assert this. | 16:38 |
| acoles | hrou yeah it was Christian I think who fixed up the header count and added the extra_header_count option | 16:38 |
| hrou | jrichli, I'm going to check now (I'll rerun with it off), it wasn't directly testing it from what I could see, we were just hitting the defined limit. I actually noticed it in many places. | 16:38 |
| jrichli | mahatic: I think you can make progress in the unit tests without rebasing. you probably want to wait on that, because i will be making changes you will want to pick up in a day anyway | 16:40 |
| mahatic | hrou: could you point me to that patch of extra_header_count? | 16:40 |
| hrou | jrichli, mahatic - 100% agreed to the above, don't bother with that ! You may run into other issues. Or if you're using a VM, clone it first : ) | 16:40 |
| mahatic | jrichli: sure, I was actually going to do this in a diff VM - which initially gave me many errors afte downloading the ecnryption code - and i didn't delete the VM yet, so I'm just curious and ready to verify quickly :) | 16:41 |
| *** Fin1te has quit IRC | 16:42 | |
| * jrichli is off to lunch | 16:43 | |
| mahatic | jrichli: enjoy your lunch! | 16:44 |
| acoles | mahatic: patch 195940 | 16:44 |
| patchbot | acoles: https://review.openstack.org/#/c/195940/ | 16:44 |
| hrou | jrichli, first time I hit: test_bad_metadata3 (test.functional.test_account.TestAccount) ... FAIL | 16:44 |
| hrou | raise HTTPException("got more than %d headers" % _MAXHEADERS)#012HTTPException: got more than 122 headers (txn: tx60e4e93309cb47b888068-0055c38e84) | 16:44 |
| mahatic | acoles: thanks! | 16:45 |
| hrou | thanks acoles ! Yep its that one, but keep in mind this didn't introduce the issue, this just added a workaround, so we should in theory have seen it before depending on your version of python | 16:45 |
| acoles | hrou sure | 16:46 |
| acoles | mahatic: hrou: also note that the doc of extra_header_count was wrong until recently, fixed here patch 209525 | 16:50 |
| patchbot | acoles: https://review.openstack.org/#/c/209525/ | 16:50 |
| acoles | just in case you get confused ;) | 16:50 |
| mahatic | acoles: cool, thanks | 16:52 |
| *** saltsa has quit IRC | 16:56 | |
| hrou | acoles, thanks ! I'm actually not convinced there still isn't an issue ; ) I found updating test.conf only didn't do the trick I needed to update the swift.conf : ) - this is just for the func tests to pass. I'll double check though. | 16:57 |
| *** saltsa has joined #openstack-swift | 16:58 | |
| *** cazino has joined #openstack-swift | 17:01 | |
| *** cazino has left #openstack-swift | 17:01 | |
| *** bapalm_ has quit IRC | 17:01 | |
| acoles | hrou: that makes sense - its the number of header from proxy->backend servers that could be exceeded by the crypto m/ware adding extra sysmeta headers. | 17:02 |
| acoles | so you'd need to increase itin swift.conf and restart | 17:02 |
| hrou | acoles, ah, that could be it, but that makes we wonder why we bother documenting it in test.conf | 17:03 |
| acoles | hrou: was just thinking the same :D | 17:04 |
| hrou | acoles, ; - ) I figure the other parameters are more dynamic in nature ? (e.g. max header size) | 17:05 |
| *** gyee has quit IRC | 17:05 | |
| mahatic | hrou: you're right, thanks for brining this up! I increased extra_header_count (directly in swift/common/constraints.py) and it worked | 17:07 |
| hrou | mahatic, by that do you mean it was failing before hand ? | 17:07 |
| hrou | mahatic, i.e. you had failing tests before making that change ? | 17:07 |
| mahatic | hrou: unit, func, probe are all giving the expected results. earlier I was getting the same headers issue | 17:07 |
| hrou | mahatic, perfect ! | 17:08 |
| hrou | mahatic, thanks ! be back in a bit. | 17:08 |
| *** petertr7_away is now known as petertr7 | 17:08 | |
| mahatic | hrou: yes, the logs give the same error too - "more than 122 headers" or the like | 17:08 |
| *** chlong has quit IRC | 17:14 | |
| openstackgerrit | Bill Huber proposed openstack/swift: Python3: do not use im_self/im_func/func_closure https://review.openstack.org/210039 | 17:16 |
| * mahatic calls it a day | 17:16 | |
| *** mahatic has quit IRC | 17:17 | |
| *** SkyRocknRoll has quit IRC | 17:20 | |
| *** dmorita has quit IRC | 17:20 | |
| *** joeljwright has quit IRC | 17:21 | |
| notmyname | good morning (again) | 17:27 |
| hrou | acoles, so in terms maximum headers, should we be more prudent then requiring someone to set extra_header_count, not sure its possible but potentially by detecting if the keymaster is in the pipeline on startup and add a fudge factor - really all this archives is passing testing - we can always hit the limit with the trivial keymaster, I'd like to understand more about this limit in httplib, I'll look into it (add an issue to the | 17:34 |
| hrou | etherpad) | 17:34 |
| *** petertr7 is now known as petertr7_away | 17:34 | |
| *** petertr7_away is now known as petertr7 | 17:35 | |
| acoles | hrou yes it definitely seems like we will need to consider either increasing the default max_header_count or documenting the need to increase the extra_header_count if using the crypto middlewares. | 17:36 |
| acoles | hrou: good catch btw! | 17:37 |
| hrou | acoles, thanks! Yea in practice this likely isn't a big deal, but at the very least we'd want to spell out that encryption implies X additional headers per user defined header (regardless of the keymaster or algo in use). | 17:43 |
| *** marcusvrn has quit IRC | 17:43 | |
| *** eandersson has quit IRC | 17:47 | |
| *** acoles is now known as acoles_ | 17:53 | |
| *** thumpba has quit IRC | 17:54 | |
| *** zul has joined #openstack-swift | 17:56 | |
| *** thumpba has joined #openstack-swift | 17:57 | |
| *** breitz has quit IRC | 17:58 | |
| *** bapalm has joined #openstack-swift | 18:04 | |
| *** thumpba_ has joined #openstack-swift | 18:14 | |
| *** nadeem has quit IRC | 18:15 | |
| *** nadeem has joined #openstack-swift | 18:16 | |
| *** thumpba has quit IRC | 18:17 | |
| jrichli | hrou mahatic acoles_: I haven't hit this yet because I hadn't rebased. I am performing rework on a patch earlier in the chain right now. | 18:21 |
| *** Fin1te has joined #openstack-swift | 18:21 | |
| *** balajir_ has joined #openstack-swift | 18:22 | |
| hrou | jrichli, thanks that makes sense ! Though is rather interesting, because it makes me wonder what the old httplib max header default was before this change, that is it mentions python 2.7.9+ but I believe you are on an earlier version ? So I wonder if the older limit was much higher, but then this change breaks it even on older versions of python. | 18:26 |
| hrou | jrichli, I say this because your using the same crypto code we are (implying we're both generating the same # of headers). | 18:26 |
| jrichli | hrou: the "extra_header_count" was added on master recently, it seems. It was added before the patch that acoles pointed out, although patch 209525 does modify that functionality. | 18:29 |
| patchbot | jrichli: https://review.openstack.org/#/c/209525/ | 18:29 |
| jrichli | hrou: I assume the test may have been modified to assert that number? But i have not verified that. All that I know is that the "bad metadata" tests pass for me right now. | 18:30 |
| hrou | jrichli, heh that second patch was just a doc fix up, alcoes mentioned both. | 18:30 |
| jrichli | what was the first? | 18:30 |
| hrou | jrichli, I don't believe the tests were intentionally stressing that actually, it was just a side effect (many failed). | 18:31 |
| hrou | jrichli, first patch that did the real change was: https://review.openstack.org/#/c/195940/ | 18:31 |
| openstackgerrit | Bill Huber proposed openstack/swift: pep8 fix: assertEquals -> assertEqual https://review.openstack.org/209976 | 18:36 |
| *** mtreinish has quit IRC | 18:37 | |
| *** mtreinish has joined #openstack-swift | 18:39 | |
| *** gyee has joined #openstack-swift | 18:40 | |
| jrichli | hrou: I haven't looked too deep, but I would guess the change that did this is the following line: | 18:41 |
| jrichli | httplib._MAXHEADERS = constraints.MAX_HEADER_COUNT | 18:41 |
| hrou | jrichli, well see that's the fix ; ) The real issue is when they introduced _MAXHEADERS into httplib in the first place, that was done through this: http://bugs.python.org/issue16037#msg182803 | 18:41 |
| hrou | specifically: http://bugs.python.org/file29201/issue16037_py27.patch | 18:42 |
| jrichli | just sayin - i think that is why the test didn't fail before, even tho the number of headers was still over | 18:43 |
| *** bill_az has quit IRC | 18:48 | |
| *** bill_az has joined #openstack-swift | 18:49 | |
| hrou | jrichli, yep! I added the issue to the etherpad. | 18:53 |
| jrichli | hrou: thx! | 19:01 |
| potato_farmer | Hi. I am trying to remove a swift storage node. I have balance, partitions, and weight down to 0. I have waited several hours between each try. But anytime I try to remove the IP, I get: "No partitions could be reassigned. Either none need to be or none can be due to min_part_hours [1]." Has anyone run into this before? | 19:02 |
| *** annegentle has joined #openstack-swift | 19:09 | |
| *** zhill_desktop has quit IRC | 19:14 | |
| *** jrichli has quit IRC | 19:15 | |
| *** breitz has joined #openstack-swift | 19:32 | |
| *** thumpba_ has quit IRC | 19:34 | |
| *** potato_farmer has quit IRC | 19:34 | |
| *** thumpba has joined #openstack-swift | 19:35 | |
| *** bill_az has quit IRC | 19:37 | |
| openstackgerrit | Eran Rom proposed openstack/swift: Add process level concurrency to container sync https://review.openstack.org/210099 | 19:37 |
| *** fifieldt_ has joined #openstack-swift | 19:40 | |
| *** amoturi has joined #openstack-swift | 19:40 | |
| *** fifieldt has quit IRC | 19:42 | |
| *** thumpba has quit IRC | 19:48 | |
| *** thumpba has joined #openstack-swift | 19:48 | |
| *** thumpba has quit IRC | 19:53 | |
| *** bill_az has joined #openstack-swift | 19:54 | |
| *** thumpba has joined #openstack-swift | 19:55 | |
| *** thumpba has quit IRC | 19:56 | |
| *** thumpba has joined #openstack-swift | 19:57 | |
| *** jamespage_ has joined #openstack-swift | 20:00 | |
| *** blmartin has quit IRC | 20:02 | |
| *** alanvitor has joined #openstack-swift | 20:06 | |
| *** _hrou_ has joined #openstack-swift | 20:09 | |
| *** robefran_ has joined #openstack-swift | 20:10 | |
| *** daemontool_ has joined #openstack-swift | 20:10 | |
| *** robefran has quit IRC | 20:11 | |
| *** ajiang has quit IRC | 20:12 | |
| openstackgerrit | Janie Richling proposed openstack/swift: FakeFooters middleware https://review.openstack.org/165517 | 20:12 |
| *** ajiang_ has joined #openstack-swift | 20:12 | |
| *** Fin1te has quit IRC | 20:12 | |
| *** hrou has quit IRC | 20:12 | |
| *** logan2 has quit IRC | 20:12 | |
| *** marzif has quit IRC | 20:12 | |
| *** logan2 has joined #openstack-swift | 20:12 | |
| *** thumpba_ has joined #openstack-swift | 20:12 | |
| *** cazino has joined #openstack-swift | 20:12 | |
| *** thumpba has quit IRC | 20:16 | |
| *** marzif has joined #openstack-swift | 20:17 | |
| *** DCWilliams_VA has joined #openstack-swift | 20:21 | |
| *** jrichli has joined #openstack-swift | 20:23 | |
| *** NM has joined #openstack-swift | 20:25 | |
| *** DCWilliams_VA has quit IRC | 20:26 | |
| *** NM1 has joined #openstack-swift | 20:30 | |
| *** NM has quit IRC | 20:32 | |
| *** tongli has quit IRC | 20:34 | |
| *** jamespage_ has quit IRC | 20:34 | |
| *** NM has joined #openstack-swift | 20:39 | |
| *** NM1 has quit IRC | 20:41 | |
| *** sanchitmalhotra has joined #openstack-swift | 20:48 | |
| *** logan2 has quit IRC | 20:48 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift-specs: PACO single-process spec. https://review.openstack.org/210117 | 20:50 |
| *** bapalm has quit IRC | 20:50 | |
| *** sanchitmalhotra1 has quit IRC | 20:50 | |
| tdasilva | ^just because we don't have enought specs to review | 20:50 |
| tdasilva | eikke, bill_az: ^^^ would like to hear your feedback | 20:51 |
| *** NM has quit IRC | 20:51 | |
| *** dustins has quit IRC | 20:59 | |
| *** thurloat is now known as thurloat_isgone | 21:02 | |
| *** petertr7 is now known as petertr7_away | 21:06 | |
| *** thumpba_ has quit IRC | 21:08 | |
| *** robefran_ has quit IRC | 21:09 | |
| pgbridge | oh that's an interesting spec | 21:15 |
| *** annegentle has quit IRC | 21:18 | |
| *** jrichli has quit IRC | 21:21 | |
| *** MVenesio has quit IRC | 21:24 | |
| openstackgerrit | Merged openstack/swift: pep8 fix: assertEquals -> assertEqual https://review.openstack.org/209652 | 21:25 |
| *** zhill has joined #openstack-swift | 21:25 | |
| *** zhill has quit IRC | 21:26 | |
| *** zhill has joined #openstack-swift | 21:27 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 21:38 | |
| *** sanchitmalhotra has quit IRC | 21:40 | |
| *** zul has quit IRC | 21:45 | |
| *** nadeem has quit IRC | 21:52 | |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: Update mock to get away from env markers https://review.openstack.org/205150 | 21:52 |
| *** robefran has joined #openstack-swift | 21:53 | |
| *** zul has joined #openstack-swift | 21:55 | |
| *** delattec has quit IRC | 21:58 | |
| *** marzif has quit IRC | 22:07 | |
| mattoliverau | Morning! | 22:17 |
| notmyname | hi | 22:17 |
| mattoliverau | Last day before flying out to Austin... I should probably think about packing :) | 22:18 |
| *** sanchitmalhotra has joined #openstack-swift | 22:22 | |
| *** sanchitmalhotra1 has quit IRC | 22:24 | |
| *** zul has quit IRC | 22:35 | |
| *** alanvitor has quit IRC | 22:38 | |
| *** _hrou_ has quit IRC | 22:41 | |
| *** wbhuber has quit IRC | 22:41 | |
| *** marzif has joined #openstack-swift | 22:42 | |
| *** chsc has quit IRC | 22:48 | |
| *** km has joined #openstack-swift | 23:00 | |
| *** jlhinson has quit IRC | 23:01 | |
| *** marzif has quit IRC | 23:01 | |
| *** logan2 has joined #openstack-swift | 23:06 | |
| *** sc68cal has quit IRC | 23:09 | |
| *** breitz has quit IRC | 23:12 | |
| *** sc68cal has joined #openstack-swift | 23:14 | |
| openstackgerrit | Merged openstack/python-swiftclient: make Connection.get_auth set url and token attributes on self https://review.openstack.org/205672 | 23:14 |
| openstackgerrit | Merged openstack/python-swiftclient: make Connection.get_auth set url and token attributes on self https://review.openstack.org/205672 | 23:14 |
| openstackgerrit | Merged openstack/swift: pep8 fix: assertEquals -> assertEqual https://review.openstack.org/209646 | 23:15 |
| openstackgerrit | Merged openstack/swift: pep8 fix: assertEquals -> assertEqual https://review.openstack.org/209649 | 23:15 |
| lcurtis | question all...I am seeing network throughput of 25mbps to my container servers but more than 10x that amount written to disk via iostat | 23:16 |
| lcurtis | what would cause that kind of write amplification to the sqlitedbs? | 23:16 |
| *** otoolee has quit IRC | 23:25 | |
| *** otoolee has joined #openstack-swift | 23:26 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 23:31 | |
| *** morganfainberg is now known as morgan_503 | 23:31 | |
| *** sanchitmalhotra has quit IRC | 23:34 | |
| notmyname | if you propose or review a security patch (in any openstack project), here's how to manage the patches: https://security.openstack.org/#how-to-propose-and-review-a-security-patch | 23:38 |
| *** breitz has joined #openstack-swift | 23:38 | |
| *** lcurtis has quit IRC | 23:40 | |
| *** zaitcev has joined #openstack-swift | 23:45 | |
| *** ChanServ sets mode: +v zaitcev | 23:45 | |
| mattoliverau | notmyname: good info, thanks :) | 23:47 |
| notmyname | and I just sent that to the ML too :-) | 23:48 |
| *** ho has joined #openstack-swift | 23:50 | |
| notmyname | ho: good morning | 23:50 |
| ho | notmyname: hello- | 23:52 |
| tdasilva | notmyname: nice blog post! | 23:55 |
| notmyname | thanks | 23:55 |
| mattoliverau | ho: morning | 23:58 |
| ho | mattoliverau: morning! | 23:59 |
| *** tsg has joined #openstack-swift | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!