| *** tqtran has quit IRC | 00:03 | |
| *** lyrrad has quit IRC | 00:24 | |
| *** Jeffrey4l has joined #openstack-swift | 00:44 | |
| *** fbo has quit IRC | 00:52 | |
| *** mhu has quit IRC | 00:54 | |
| *** siva_krish has quit IRC | 00:59 | |
| *** fbo has joined #openstack-swift | 01:03 | |
| *** Suyash has quit IRC | 01:05 | |
| *** klrmn has quit IRC | 01:11 | |
| *** mhu has joined #openstack-swift | 01:12 | |
| *** dmorita has quit IRC | 01:12 | |
| *** awelleck has joined #openstack-swift | 01:21 | |
| *** StraubTW has joined #openstack-swift | 01:22 | |
| *** StraubTW has quit IRC | 01:23 | |
| *** cdelatte has quit IRC | 01:38 | |
| *** Suyash has joined #openstack-swift | 01:45 | |
| *** siva_krish has joined #openstack-swift | 01:53 | |
| *** siva_krish has quit IRC | 01:59 | |
| *** tqtran has joined #openstack-swift | 02:01 | |
| *** zul has quit IRC | 02:03 | |
| *** zul has joined #openstack-swift | 02:04 | |
| *** tqtran has quit IRC | 02:05 | |
| *** klrmn has joined #openstack-swift | 02:14 | |
| *** awelleck has quit IRC | 02:21 | |
| *** zul has quit IRC | 02:23 | |
| *** dmorita has joined #openstack-swift | 02:24 | |
| *** greghaynes has quit IRC | 02:27 | |
| *** dmorita has quit IRC | 02:28 | |
| *** _JZ_ has quit IRC | 02:37 | |
| *** greghaynes has joined #openstack-swift | 02:39 | |
| *** awelleck has joined #openstack-swift | 02:51 | |
| *** awelleck has quit IRC | 02:51 | |
| *** stevemar has joined #openstack-swift | 03:36 | |
| *** links has joined #openstack-swift | 03:51 | |
| *** SkyRocknRoll has joined #openstack-swift | 03:52 | |
| *** ppai has joined #openstack-swift | 03:57 | |
| *** psachin has joined #openstack-swift | 04:38 | |
| *** klrmn1 has joined #openstack-swift | 04:42 | |
| *** klrmn has quit IRC | 04:43 | |
| openstackgerrit | Pete Zaitcev proposed openstack/swift: Patch through the error from NewReplicator https://review.openstack.org/333107 | 04:44 |
|---|---|---|
| *** Suyash has quit IRC | 04:49 | |
| openstackgerrit | Pete Zaitcev proposed openstack/swift: Add a global fix for anything that involves GetHashPrefixAndSuffix https://review.openstack.org/333108 | 04:51 |
| *** zaitcev has quit IRC | 05:21 | |
| *** ChubYann has quit IRC | 05:26 | |
| *** dmorita has joined #openstack-swift | 05:27 | |
| *** dmorita has quit IRC | 05:32 | |
| *** tqtran has joined #openstack-swift | 06:02 | |
| *** tqtran has quit IRC | 06:06 | |
| *** rcernin has joined #openstack-swift | 06:15 | |
| *** jmccarthy has quit IRC | 06:42 | |
| *** jmccarthy has joined #openstack-swift | 06:44 | |
| *** jamielennox is now known as jamielennox|away | 06:48 | |
| *** oshritf has joined #openstack-swift | 06:54 | |
| *** rcernin has quit IRC | 07:02 | |
| *** tesseract- has joined #openstack-swift | 07:13 | |
| *** tesseract- has quit IRC | 07:14 | |
| *** tesseract- has joined #openstack-swift | 07:14 | |
| *** rcernin has joined #openstack-swift | 07:17 | |
| *** furlongm_ has joined #openstack-swift | 07:22 | |
| *** furlongm has quit IRC | 07:22 | |
| *** rledisez has joined #openstack-swift | 07:25 | |
| *** furlongm has joined #openstack-swift | 07:26 | |
| *** furlongm_ has quit IRC | 07:27 | |
| *** hseipp has joined #openstack-swift | 07:33 | |
| *** furlongm has quit IRC | 07:35 | |
| *** furlongm has joined #openstack-swift | 07:35 | |
| *** SkyRocknRoll_ has joined #openstack-swift | 07:37 | |
| *** SkyRocknRoll_ has quit IRC | 07:38 | |
| *** oshritf has quit IRC | 07:39 | |
| *** furlongm has quit IRC | 07:42 | |
| *** geaaru has joined #openstack-swift | 07:49 | |
| *** klrmn1 has quit IRC | 07:57 | |
| *** baojg has joined #openstack-swift | 08:05 | |
| *** geaaru has quit IRC | 08:16 | |
| *** mvk_ has quit IRC | 08:21 | |
| *** geaaru has joined #openstack-swift | 08:28 | |
| *** hosanai has joined #openstack-swift | 08:30 | |
| *** ChanServ sets mode: +v hosanai | 08:30 | |
| *** daemontool has joined #openstack-swift | 08:31 | |
| *** Jeffrey4l has quit IRC | 08:47 | |
| *** rcernin has quit IRC | 08:50 | |
| *** Jeffrey4l has joined #openstack-swift | 09:00 | |
| *** rcernin has joined #openstack-swift | 09:02 | |
| *** ouchkernel has quit IRC | 09:03 | |
| *** tqtran has joined #openstack-swift | 09:03 | |
| *** ouchkernel has joined #openstack-swift | 09:07 | |
| *** tqtran has quit IRC | 09:08 | |
| *** dmk0202 has joined #openstack-swift | 09:09 | |
| *** mvk_ has joined #openstack-swift | 09:11 | |
| *** silor has joined #openstack-swift | 09:11 | |
| *** silor1 has joined #openstack-swift | 09:16 | |
| *** silor has quit IRC | 09:18 | |
| *** silor1 is now known as silor | 09:18 | |
| *** zul has joined #openstack-swift | 09:23 | |
| *** mvk_ has quit IRC | 09:24 | |
| *** baojg has quit IRC | 09:26 | |
| *** Jeffrey4l has quit IRC | 09:28 | |
| *** dmorita has joined #openstack-swift | 09:28 | |
| *** dmorita has quit IRC | 09:33 | |
| *** kei_yama has quit IRC | 09:38 | |
| *** cbartz has joined #openstack-swift | 09:38 | |
| *** acoles_ is now known as acoles | 09:41 | |
| *** Jeffrey4l has joined #openstack-swift | 09:42 | |
| *** mvk_ has joined #openstack-swift | 09:43 | |
| *** jamielennox|away is now known as jamielennox | 09:43 | |
| *** pcaruana has joined #openstack-swift | 09:51 | |
| *** daemontool has quit IRC | 10:00 | |
| *** foexle has joined #openstack-swift | 10:01 | |
| *** zul has quit IRC | 10:23 | |
| *** dmk0202 has quit IRC | 10:29 | |
| *** zul has joined #openstack-swift | 10:35 | |
| *** zul has quit IRC | 10:38 | |
| *** zul has joined #openstack-swift | 10:52 | |
| *** jistr is now known as jistr|mtg | 10:56 | |
| *** zul has quit IRC | 11:16 | |
| *** zul has joined #openstack-swift | 11:17 | |
| *** ekarlso has quit IRC | 11:26 | |
| openstackgerrit | oshritf proposed openstack/swift: Add thread level concurrency to container sync https://review.openstack.org/225338 | 11:26 |
| *** daemontool has joined #openstack-swift | 11:28 | |
| *** dmk0202 has joined #openstack-swift | 11:33 | |
| openstackgerrit | oshritf proposed openstack/swift: Add process level concurrency to container sync https://review.openstack.org/332985 | 11:50 |
| *** ekarlso has joined #openstack-swift | 11:51 | |
| *** daemontool_ has joined #openstack-swift | 11:54 | |
| *** daemontool has quit IRC | 11:55 | |
| openstackgerrit | Victor Stinner proposed openstack/swift: Python 3: Fix basestring, long and StringIO https://review.openstack.org/333297 | 12:02 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Python 3: dict.iteritems() and dict.itervalues() https://review.openstack.org/333298 | 12:02 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Python 3: fix urllib import https://review.openstack.org/333299 | 12:02 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Python 3: fix usage of reload() https://review.openstack.org/333300 | 12:02 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Python 3: fix ConfigParser and httplib imports https://review.openstack.org/333301 | 12:02 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Python 3: Fix usage of dict methods https://review.openstack.org/333302 | 12:02 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Python 3: fix dict.values()[0] https://review.openstack.org/333303 | 12:02 |
| *** dmorita has joined #openstack-swift | 12:09 | |
| *** dmorita has quit IRC | 12:14 | |
| *** manous has joined #openstack-swift | 12:35 | |
| *** SkyRocknRoll has quit IRC | 12:36 | |
| *** ametts has joined #openstack-swift | 12:43 | |
| *** tongli has joined #openstack-swift | 12:48 | |
| *** McMurlock1 has joined #openstack-swift | 12:49 | |
| *** janonymous has quit IRC | 12:51 | |
| openstackgerrit | Christopher Bartz proposed openstack/swift: Preserve query params in tempurl https://review.openstack.org/333331 | 12:57 |
| *** hosanai has quit IRC | 12:57 | |
| *** pauloewerton has joined #openstack-swift | 12:58 | |
| *** manous_ has joined #openstack-swift | 13:09 | |
| *** manous has quit IRC | 13:11 | |
| *** StraubTW has joined #openstack-swift | 13:14 | |
| tdasilva | good morning | 13:23 |
| *** jistr|mtg is now known as jistr | 13:31 | |
| *** catintheroof has joined #openstack-swift | 13:32 | |
| ntata | good morning | 13:35 |
| *** foexle has quit IRC | 13:46 | |
| *** zul_ has joined #openstack-swift | 13:59 | |
| *** zul has quit IRC | 14:00 | |
| *** mwheckmann has joined #openstack-swift | 14:00 | |
| *** ppai has quit IRC | 14:00 | |
| *** zul_ has quit IRC | 14:01 | |
| *** zul has joined #openstack-swift | 14:02 | |
| *** manous_ has quit IRC | 14:04 | |
| *** manous_ has joined #openstack-swift | 14:18 | |
| onovy | hi, https://review.openstack.org/#/c/323874/ and https://review.openstack.org/#/c/323877/ anyone? ;) | 14:22 |
| patchbot | onovy: patch 323874 - pyeclib - Change liberasure URL to new one | 14:22 |
| patchbot | onovy: patch 323877 - pyeclib - Bump liberasurecode dependency to 1.2.0 version | 14:22 |
| onovy | and https://review.openstack.org/#/c/323819/ :) | 14:22 |
| patchbot | onovy: patch 323819 - liberasurecode - Changed homepage to new one | 14:22 |
| *** rfeusi has joined #openstack-swift | 14:24 | |
| *** foexle has joined #openstack-swift | 14:26 | |
| rfeusi | I need help for a minimal openstack swift architecture. I have two datacenters next to each others (only a steet is between them, so latency is very small). | 14:27 |
| *** jistr is now known as jistr|mtg | 14:38 | |
| *** links has quit IRC | 14:41 | |
| *** pcaruana has quit IRC | 14:41 | |
| onovy | rfeusi: hi, we are using geocluster for this, so two regions and two replicas each | 14:48 |
| *** foexle has quit IRC | 14:49 | |
| onovy | but if latency is really small, and throughtput betwean them is not issue, you can make it as one region too | 14:49 |
| *** manous_ has quit IRC | 14:56 | |
| *** siva_krish has joined #openstack-swift | 14:57 | |
| *** mvk_ has quit IRC | 15:02 | |
| *** amit213 has quit IRC | 15:04 | |
| *** tqtran has joined #openstack-swift | 15:06 | |
| *** tqtran has quit IRC | 15:10 | |
| notmyname | good morning | 15:14 |
| *** sheel has quit IRC | 15:15 | |
| *** siva_krish has left #openstack-swift | 15:17 | |
| *** dmorita has joined #openstack-swift | 15:19 | |
| *** haypo has joined #openstack-swift | 15:33 | |
| *** Suyash has joined #openstack-swift | 15:33 | |
| *** dmk0202 has quit IRC | 15:46 | |
| *** psachin has quit IRC | 15:47 | |
| *** lyrrad has joined #openstack-swift | 15:50 | |
| *** arch-nemesis has joined #openstack-swift | 15:52 | |
| timburke | good morning | 15:52 |
| *** diogogmt has joined #openstack-swift | 16:03 | |
| *** chsc has joined #openstack-swift | 16:09 | |
| *** rledisez has quit IRC | 16:17 | |
| *** tesseract- has quit IRC | 16:17 | |
| *** baojg has joined #openstack-swift | 16:18 | |
| *** dosaboy has quit IRC | 16:18 | |
| *** Suyash has quit IRC | 16:18 | |
| *** sheel has joined #openstack-swift | 16:22 | |
| notmyname | jrichli: in keymaster.py, we're using sha256 for the hmac. what would you think about changing that to sha384? | 16:32 |
| notmyname | I just made the change locally, and I'm about to run tests against it | 16:33 |
| *** Suyash has joined #openstack-swift | 16:36 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:36 | |
| *** dosaboy has joined #openstack-swift | 16:39 | |
| *** jistr|mtg is now known as jistr | 16:42 | |
| notmyname | well that didn't work | 16:42 |
| jrichli | notmyname: sounds good. just wondering if there was a particular compliance issue that influenced that change? | 16:42 |
| jrichli | oh :-/ | 16:42 |
| notmyname | and neither did the second thign | 16:42 |
| *** SkyRocknRoll has quit IRC | 16:42 | |
| notmyname | jrichli: I tried https://gist.github.com/notmyname/6b29fb06a7d32c9ee5911896265ece8b and I get a client disconnect (499) from the object servers | 16:43 |
| notmyname | the reason I'm curious is for "checkbox compliance" | 16:43 |
| notmyname | basically, when looking at some US government standards, AES256 is good for up to TOP SECRET, but SHA256 is only good to SECRET. SHA384 is also ok for TOP SECRET | 16:44 |
| *** silor has quit IRC | 16:44 | |
| jrichli | but is that for encryption, or key derivation? | 16:44 |
| jrichli | the use case matters | 16:45 |
| notmyname | checkbox compliance ;-) | 16:45 |
| jrichli | id have to see the description of the checkbox item | 16:45 |
| notmyname | from that thing I linked yesterday...looking for the link again | 16:45 |
| jrichli | i have that link. | 16:45 |
| jrichli | ill take a look | 16:45 |
| jrichli | btw - we are needing a 256 bit key because of our encryption | 16:46 |
| notmyname | yeah, the suite B list on https://www.cnss.gov/CNSS/openDoc.cfm?lZsorLEWqomRp4GEC0ztzA== | 16:46 |
| notmyname | but seriously. that's my whole thinking. someone asked if the algorithms used in swift are on the suite B list, and that's it | 16:47 |
| notmyname | so I thought it might be a relatively simple change. turns out it's probably not :-) | 16:48 |
| jrichli | "the algorithms used" usually means for the encyrption part | 16:48 |
| notmyname | yeah, but sha isn't an encryption algorithm | 16:48 |
| notmyname | and it's on the list | 16:48 |
| jrichli | but deriving ivs or keys is a whole different use case. the checkbox would be specific if it is addressing key derivation | 16:48 |
| jrichli | 256-bit keys is what we decided on | 16:48 |
| jrichli | if you try to run with a longer key, crypto will complain | 16:49 |
| notmyname | yes I know (now) ;-) | 16:49 |
| notmyname | but yeah, I see what you mean now | 16:50 |
| notmyname | and I because of what you said, I understand the question better. thanks | 16:50 |
| jrichli | so if AES with 256 bit keys is on the list (and i thought it was) than we dont need longer keys | 16:50 |
| jrichli | ok, cool | 16:50 |
| *** cbartz has left #openstack-swift | 16:55 | |
| *** klrmn has joined #openstack-swift | 17:03 | |
| *** baojg has quit IRC | 17:05 | |
| *** daemontool_ has quit IRC | 17:06 | |
| timburke | notmyname: do you have a title for that document? every time i try to follow your link, i get "We're sorry, but we're having trouble opening this document because we did not recognize your session variable." blah blah "Not all pages in this site can be bookmarked." | 17:07 |
| notmyname | timburke: it's the CNSSP 15 on https://www.cnss.gov/CNSS/issuances/Policies.cfm | 17:07 |
| timburke | thanks | 17:08 |
| jrichli | FWIW: this is what cca has said about our key derivation security: "Using HMAC for key derivation is sound and fits. The HKDF function serves a different scenario, more appropriate for a protocol context. Here we are in the scenario of NIST SP 800-108, and its KDF in counter mode with HMAC (sec. 5.1) is equivalent in security to what we use here. It is of critical importance that 'encryption_root_secret' has enough entropy (256 bits o | 17:09 |
| jrichli | r more)." | 17:09 |
| *** McMurlock1 has quit IRC | 17:09 | |
| jrichli | of course, this doesn't mention the hashing alg used within the HMAC call, but he knows we are using 256 bit keys. | 17:11 |
| *** Jeffrey4l has quit IRC | 17:11 | |
| notmyname | ok | 17:11 |
| notmyname | I just learned about https://blake2.net. It looks really interesting as an alternative to where md5 is used in swift today (yes, I realize that would be a HUGE change) | 17:12 |
| notmyname | just as fast or faster than md5, and it has some interesting options to have variable digest sizes depending on the use case | 17:12 |
| notmyname | and the FAQ even mentions that it's good for stuff like OpenStack Swift! | 17:13 |
| *** tqtran has joined #openstack-swift | 17:13 | |
| jrichli | interesting | 17:14 |
| timburke | i wonder if the cryptography package would support 3AES with a 512 bit key (say, coming out of an HMAC-SHA512), similar to 3DES...not that i think we necessarily need or want to. just idle curiosity | 17:15 |
| *** ChubYann has joined #openstack-swift | 17:17 | |
| timburke | oh hey, the FAQ even mentions Swift by name! | 17:18 |
| notmyname | timburke: yeah. cool, huh? | 17:18 |
| jrichli | notmyname: the problem about researching the 'goodness' of our use of sha256 here is that typically HMAC is used for message authentication. But we are not using it for that in this case. | 17:20 |
| jrichli | but the issue is collisions - and yes, we want to not repeat keys if we can help it | 17:20 |
| jrichli | its a lot simpler if you can just say you generate random keys :-) | 17:21 |
| *** raildo is now known as raildo-afk | 17:22 | |
| jrichli | similar issue with discussions of md5 use in swift. we dont use it as an HMAC. so what everybody talks about doesnt apply | 17:22 |
| notmyname | right :-) | 17:22 |
| timburke | speaking of random keys... do we have any data regarding the point at which having sufficient entropy becomes a bottleneck? | 17:24 |
| jrichli | a bottleneck to what exactly? | 17:25 |
| *** hseipp has quit IRC | 17:26 | |
| timburke | if i PUT a whole bunch of tiny (<32 bytes, say) objects, the proxies will be getting 32 bytes worth of random key material for each. under enough load, i could see the proxy needing to either wait for more random data to be generated or accept lower-quality data | 17:34 |
| *** rcernin has quit IRC | 17:38 | |
| timburke | random experimenting would seem to indicate this is unlikely to be a problem. i think the standard handling-a-PUT overhead will far outweigh waiting on 32 bytes of entropy | 17:44 |
| notmyname | timburke: that's when you fork over $15 and plug one of these into the server http://altusmetrum.org/ChaosKey/ | 17:46 |
| timburke | notmyname: knowing when you need one is good, though | 17:48 |
| notmyname | sure. I guess "when /dev/urandom is slow" probably isn't a good metric ;0( | 17:49 |
| notmyname | or whatever that smiley thing is supposed to be | 17:49 |
| *** mwheckmann has quit IRC | 17:51 | |
| timburke | notmyname: looks like v0.4 has a red LED! that's when you *know* you've made it http://git.gag.com/?p=hw/usbtrng | 17:53 |
| notmyname | whoa! | 17:53 |
| notmyname | timburke: I'm sure v0.5 will have a blue LED | 17:54 |
| timburke | notmyname: too late! maybe v0.6 http://git.gag.com/?p=hw/usbtrng;a=commit;h=531f5ef | 17:57 |
| notmyname | I remember once upon a time when redbo got fed up with how long the proxy server code was and refactored it into all the different proxy controllers. and now I'm scrolling through 2300+ lines of just the proxy object controller thinking it should be split up | 17:59 |
| notmyname | what with all the different putters and mimecontrollers | 18:00 |
| *** McMurlock1 has joined #openstack-swift | 18:05 | |
| *** pauloewerton has quit IRC | 18:11 | |
| *** geaaru has quit IRC | 18:14 | |
| * jrichli is back from lunch. reading now | 18:14 | |
| jrichli | timburke: so you are talking about a scenario where somebody were to use random keys in swift? Or are you talking about the possible strain of the random iv generation for our current scheme? | 18:20 |
| *** zul has quit IRC | 18:20 | |
| timburke | jrichli: current scheme. but like i said, it doesn't seem to add much overhead; i'm not worried | 18:21 |
| jrichli | timburke: i am probably missing something. we dont use random for anything except the ivs. the item of sufficient entropy that christian was referring to was the root key that is fixed and supplied in config. | 18:22 |
| timburke | yes, and that seems fine. i like our requirements on the root key | 18:24 |
| timburke | i'm thinking of the random ivs and body keys that are generated per PUT | 18:24 |
| *** zul has joined #openstack-swift | 18:24 | |
| jrichli | timburke: oh, right. we switched to random key for body sorta late in the game (when wrapping was introduced). i had it my mind we were still using the derived key. | 18:27 |
| timburke | with max meta around 90, we're talking about only <1.5k worth of random data per request | 18:28 |
| tdasilva | notmyname: will the crypto work be merged to master as one single commit or multiple commits? | 18:29 |
| timburke | tdasilva: the plan is to have multiple commits on the crypto-review branch which all come into master as a single merge commit | 18:29 |
| notmyname | tdasilva: when we all like the crypto-review chain, we'll propose one merge commit to land on master | 18:29 |
| *** siva_krish has joined #openstack-swift | 18:30 | |
| tdasilva | notmyname, timburke: ok, was just thinking earlier today if it would make sense to start merging separate patches as we are "done" with it...the first patch already has 3 +2s. I was wondering if that would make things easier in terms of maintaining the chain | 18:31 |
| notmyname | yeah, perhaps. actually, I want to leave that up to acoles and whatever local tools he has to manage it. but yeah, I'd be fine with landing the first one on the crypto-review branch now. I'd also be ok with leaving it as-is | 18:32 |
| timburke | leaving it as-is has the benefit of letting us go back and change it if we somehow realized while looking at patch 5 (say) that something else should have been done in patch 1. at the same time, maybe the fact that we didn't realize it until patch 5 demonstrates that the change logically belongs in patch 5 anyway... | 18:34 |
| patchbot | timburke: https://review.openstack.org/#/c/5/ - openstack-infra/system-config - Fix problem with jenkins known_hosts url. (MERGED) | 18:34 |
| patchbot | timburke: https://review.openstack.org/#/c/1/ - openstack-infra/system-config - Add puppet module for ssh that installs an sshd_co... (MERGED) | 18:34 |
| patchbot | timburke: https://review.openstack.org/#/c/5/ - openstack-infra/system-config - Fix problem with jenkins known_hosts url. (MERGED) | 18:34 |
| patchbot | timburke: https://review.openstack.org/#/c/5/ - openstack-infra/system-config - Fix problem with jenkins known_hosts url. (MERGED) | 18:34 |
| timburke | :-( | 18:35 |
| notmyname | tdasilva: don't poke happy fun patchbot | 18:35 |
| notmyname | timburke: ^ | 18:35 |
| tdasilva | lol | 18:36 |
| *** mlanner has quit IRC | 18:37 | |
| *** McMurlock1 has quit IRC | 18:37 | |
| jrichli | wow. threre really patch numbers that low | 18:37 |
| acoles | I can't resist.... patch 0 | 18:38 |
| patchbot | acoles: https://review.openstack.org/#/c/0/ | 18:38 |
| tdasilva | timburke: yeah, I thought of that too, but I think that could also be easily fixed. I guess I was striving for a couple of things. First really just perception. If we get to next Friday and there's only two or three patches in the chain to merge, I'd feel better. Also, it's really the last few patches that have the crypto work. Merging earlier patches as they are "done" might allow for a "softer soft freeze" on master | 18:39 |
| tdasilva | and we don't have to worry so much about conflicts | 18:39 |
| *** mlanner has joined #openstack-swift | 18:40 | |
| *** tongli has quit IRC | 18:40 | |
| acoles | tdasilva: are you suggesting merging the earlier patches to master? otherwise I think it makes no difference wrt conflicts if they are merged on crypto-review or not | 18:41 |
| acoles | but I will consider whether it could make life easier for me or reviewers in any other way | 18:42 |
| tdasilva | acoles: yes, merge to master | 18:43 |
| tdasilva | acoles: what got me thinking about this was the convo yesterday about the documentation and the sync work that is being held up | 18:43 |
| acoles | oic. | 18:43 |
| jrichli | I think tdasilva has a good idea. I think at the rate we are going, one more week isn't going to be enough (I hate to say) | 18:43 |
| timburke | as i recall, notmyname seemed opposed to it, but his opinion may have changed | 18:44 |
| timburke | oh yeah, i meant to mention this in the meeting: i've been using a dash like https://goo.gl/f9gMj4 - other reviewers may find it useful. puts patches where you have a draft at the top, and categorizes all of them based on whether you left a positive or negative (or no) score | 18:44 |
| notmyname | no, I don't want to merge to master | 18:45 |
| acoles | the sync work shouldn't conflict, I wonder if actually it was in conflict with master and hence crypto-review | 18:45 |
| notmyname | but landing on crypto-review is ok | 18:45 |
| acoles | the one advantage of landing on crypto-review is that if we want to go back and change stuff that landed we *may* be able to do it with a fresh patch at the top of the chain, and that avoids lots of rebase noise for the other patches. the disadvantage is that it opens the flood gates for adding patches to the top of the chain ;) | 18:47 |
| notmyname | I'd be fine with landing on crypto-review, but I'd have the same concern as timburke mentioned | 18:48 |
| notmyname | and I don't think the docs works should likely be held up for much longer. I do think we have a pretty good shot at landing this in the next 7 days | 18:49 |
| jrichli | yes, i see that point too. | 18:49 |
| acoles | as far as my process is concerned, I don't yet have a problem with managing an un-merged chain | 18:49 |
| acoles | notmyname: +1 - done in 7 days or I -2 all the patches and ragequit ;) | 18:50 |
| notmyname | :-) | 18:51 |
| jrichli | acoles: dont even joke about that! | 18:51 |
| notmyname | we'll have the alexit | 18:51 |
| acoles | oh, you just reminded me I need to vote! | 18:52 |
| notmyname | do you have to go to a polling station, or can you do it online? | 18:53 |
| notmyname | ahale and joeljwright need to go vote too :-) | 18:53 |
| acoles | online ? (!*"$^ that'll be the day! believe it or not we use paper and pencil and there is no id check, you just state your name/address and get to vote. | 18:54 |
| notmyname | same here (when it's not the closed-source electronic machines) | 18:55 |
| clarkb | oregon and washington's mail in voting system is really nice | 18:57 |
| *** amit213 has joined #openstack-swift | 19:00 | |
| redbo | In texas, you have to show your gun to vote. | 19:00 |
| *** sheel has quit IRC | 19:05 | |
| notmyname | jrichli: on patch 328205 do you think we'll drop the x-backend-container-update-override-* at some point later? | 19:11 |
| patchbot | notmyname: https://review.openstack.org/#/c/328205/ - swift (feature/crypto-review) - Make container update override headers persistent | 19:11 |
| notmyname | you know, as much as we drop anything in swift | 19:11 |
| notmyname | jrichli: are you working with thai tran? | 19:14 |
| jrichli | notmyname: to some extent yes, I with thai. as for backend, i will have to look at some notes i had. in a call now : will get back to you | 19:17 |
| timburke | notmyname: the ones sent for EC support? i don't think so. the object server needs to account for it regardless, due to our upgrade-storage-nodes-then-upgrade-proxy-nodes recommendation. and since the object server needs to handle it anyway, we may as well keep sending it; then even if a proxy server upgrades before an object server, this particular thing will still work | 19:19 |
| notmyname | /lunch | 19:22 |
| *** htruta is now known as henrique | 19:22 | |
| *** henrique is now known as Guest78419 | 19:23 | |
| *** Guest78419 is now known as htruta | 19:23 | |
| *** siva_krish has quit IRC | 19:25 | |
| *** welldannit has joined #openstack-swift | 19:26 | |
| *** chsc has quit IRC | 19:35 | |
| *** haypo has left #openstack-swift | 19:36 | |
| acoles | notmyname: +1 what timburke said, don't see us dropping the old-style overrides for EC | 19:46 |
| *** acoles is now known as acoles_ | 19:47 | |
| *** zaitcev has joined #openstack-swift | 19:50 | |
| *** ChanServ sets mode: +v zaitcev | 19:50 | |
| openstackgerrit | Tim Burke proposed openstack/swift: MockMemcached cleanup https://review.openstack.org/333554 | 19:56 |
| openstackgerrit | Tim Burke proposed openstack/swift: Tighten memcached tests https://review.openstack.org/333555 | 19:56 |
| ahale | hehe , do you have to shoot the polling card too redbo ? | 20:00 |
| redbo | Yeah, at 20 yards | 20:00 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Python 3: Fix usage of dict methods https://review.openstack.org/333302 | 20:02 |
| openstackgerrit | Victor Stinner proposed openstack/swift: Python 3: fix dict.values()[0] https://review.openstack.org/333303 | 20:02 |
| *** siva_krish has joined #openstack-swift | 20:06 | |
| *** chsc has joined #openstack-swift | 20:11 | |
| *** gyee has joined #openstack-swift | 20:16 | |
| zaitcev | My wife has a Walther Q5 that can easily do that. | 20:16 |
| jrichli | notmyname: btw, thai tran is tqtran | 20:24 |
| *** zul has quit IRC | 20:27 | |
| *** fbo has quit IRC | 20:27 | |
| *** hugokuo has quit IRC | 20:27 | |
| *** rsFF- has quit IRC | 20:27 | |
| *** remix_tj has quit IRC | 20:27 | |
| *** Anticimex has quit IRC | 20:27 | |
| *** MooingLemur has quit IRC | 20:27 | |
| *** peterlisak has quit IRC | 20:27 | |
| *** zigo has quit IRC | 20:27 | |
| *** ntata has quit IRC | 20:27 | |
| *** onovy has quit IRC | 20:27 | |
| *** _fortis has quit IRC | 20:27 | |
| *** CrackerJ- has quit IRC | 20:27 | |
| *** zacksh_ has quit IRC | 20:27 | |
| *** mmotiani has quit IRC | 20:27 | |
| *** j_king has quit IRC | 20:27 | |
| *** jamielennox has quit IRC | 20:27 | |
| *** mattoliverau has quit IRC | 20:27 | |
| *** mmotiani_ is now known as mmotiani | 20:27 | |
| *** j_king has joined #openstack-swift | 20:27 | |
| *** matt6434 has joined #openstack-swift | 20:27 | |
| *** rsFF has joined #openstack-swift | 20:27 | |
| *** onovy has joined #openstack-swift | 20:27 | |
| *** CrackerJackMack has joined #openstack-swift | 20:27 | |
| *** Anticimex has joined #openstack-swift | 20:27 | |
| *** hugokuo has joined #openstack-swift | 20:27 | |
| *** peterlisak has joined #openstack-swift | 20:27 | |
| *** MooingLemur has joined #openstack-swift | 20:27 | |
| *** remix_tj has joined #openstack-swift | 20:27 | |
| *** mmotiani_ has joined #openstack-swift | 20:28 | |
| *** zul has joined #openstack-swift | 20:28 | |
| *** zacksh has joined #openstack-swift | 20:28 | |
| *** zigo has joined #openstack-swift | 20:29 | |
| *** ouchkernel has quit IRC | 20:29 | |
| *** fbo has joined #openstack-swift | 20:30 | |
| *** McMurlock1 has joined #openstack-swift | 20:36 | |
| *** ouchkernel has joined #openstack-swift | 20:36 | |
| *** cebreidian has joined #openstack-swift | 20:36 | |
| *** ntata has joined #openstack-swift | 20:37 | |
| *** jamielennox has joined #openstack-swift | 20:38 | |
| *** dmorita has quit IRC | 20:43 | |
| *** dmorita has joined #openstack-swift | 20:44 | |
| *** adu has joined #openstack-swift | 20:50 | |
| *** mmcardle has quit IRC | 20:50 | |
| *** mvk_ has joined #openstack-swift | 21:02 | |
| *** nadeem has joined #openstack-swift | 21:07 | |
| *** foexle has joined #openstack-swift | 21:07 | |
| *** ozialien10 has quit IRC | 21:08 | |
| *** McMurlock1 has quit IRC | 21:17 | |
| timburke | zaitcev: you may also like patch 333554 | 21:19 |
| patchbot | timburke: https://review.openstack.org/#/c/333554/ - swift - MockMemcached cleanup | 21:19 |
| *** catintheroof has quit IRC | 21:21 | |
| *** chsc has quit IRC | 21:23 | |
| *** adu has quit IRC | 21:26 | |
| zaitcev | if new_val[0] == '-': # ie, val is negative | 21:28 |
| timburke | it felt less stupid than eval'ing str(int(current) + int(value)) twice. not by much, i'll admit... | 21:30 |
| timburke | er, minus, not plus. you know what i mean | 21:31 |
| zaitcev | It just seemed to me much too obvious. | 21:32 |
| timburke | i'm always happy to remove comments :-) | 21:32 |
| zaitcev | I'll check if list+list still works in py3 after tea | 21:33 |
| zaitcev | actually looks like tuple+tuple here self.cache[key][:2] + (new_val, ) | 21:33 |
| zaitcev | else it's .append() or something, I dunn | 21:34 |
| notmyname | google has a nice little readout https://www.google.com/#q=brexit+results | 21:42 |
| timburke | zaitcev: should be fine. py3 won't join two views (like from dict.keys() or dict.values()), but lists and tuples still work | 21:47 |
| zaitcev | notmyname: You can tell it's going to be 51/49 for Remain without any Google. YOUGOV said so, it is so. | 21:55 |
| *** chsc has joined #openstack-swift | 21:56 | |
| *** chsc has joined #openstack-swift | 21:56 | |
| *** rfeusi has quit IRC | 22:01 | |
| *** rfeusi has joined #openstack-swift | 22:02 | |
| *** ametts has quit IRC | 22:02 | |
| *** nadeem has quit IRC | 22:17 | |
| *** matt6434 is now known as mattoliverau | 22:18 | |
| *** ChanServ sets mode: +v mattoliverau | 22:18 | |
| mattoliverau | morning | 22:18 |
| *** ouchkernel has quit IRC | 22:19 | |
| *** acoles_ is now known as acoles | 22:24 | |
| *** McMurlock1 has joined #openstack-swift | 22:27 | |
| acoles | notmyname: this is the best brexit indicator http://www.xe.com/currencycharts/?from=GBP&to=USD&view=12h | 22:27 |
| notmyname | :-) | 22:28 |
| *** StraubTW has quit IRC | 22:30 | |
| *** diogogmt has quit IRC | 22:33 | |
| *** chsc has quit IRC | 22:36 | |
| *** chsc has joined #openstack-swift | 22:41 | |
| *** chsc has joined #openstack-swift | 22:41 | |
| *** natarej has quit IRC | 22:43 | |
| *** diogogmt has joined #openstack-swift | 22:44 | |
| *** natarej has joined #openstack-swift | 22:44 | |
| openstackgerrit | Merged openstack/swift: Patch through the error from NewReplicator https://review.openstack.org/333107 | 22:49 |
| clayg | notmyname: ok, how do you crypto? | 22:50 |
| timburke | yay! | 22:51 |
| notmyname | clayg: the patches are organized really well. I started with the whole-view (ie start at the end) to understand the pieces, and now I'm going through from patch 0 to 6 | 22:51 |
| notmyname | the commit messages are really good at explaining the need and what's being done. starts with a lot of refactoring, then adds the new functionality, then docs | 22:52 |
| acoles | oh no | 22:52 |
| timburke | haha | 22:52 |
| * acoles runs for cover | 22:52 | |
| clayg | timburke: how did patch 333554 happen? did you see something blow up? | 22:53 |
| patchbot | clayg: https://review.openstack.org/#/c/333554/ - swift - MockMemcached cleanup | 22:53 |
| timburke | clayg: i saw patch 333303 and decided it didn't do enough to improve the situation | 22:53 |
| patchbot | timburke: https://review.openstack.org/#/c/333303/ - swift - Python 3: fix dict.values()[0] | 22:53 |
| acoles | clayg: patch 328209 might be a good place to start, before diving into the crypto code in patch 328208. 328203 - 328207 are supporting changes. | 22:55 |
| patchbot | acoles: https://review.openstack.org/#/c/328209/ - swift (feature/crypto-review) - Add encryption overview doc | 22:55 |
| patchbot | acoles: https://review.openstack.org/#/c/328208/ - swift (feature/crypto-review) - Enable object body and metadata encryption | 22:55 |
| clayg | oh yay! py3 patches are back! yay! | 22:55 |
| timburke | clayg: a relevant thread for that starts at http://lists.openstack.org/pipermail/openstack-dev/2016-June/097930.html - though i'm not sure how much you'd care to read it | 22:56 |
| clayg | acoles: nice work on the rebase - chain looks solid | 22:59 |
| timburke | acoles: good point on the derived iv! so if we use the object key for object-server etags...could we get away with using the same iv for all objects? hmm... | 23:01 |
| acoles | timburke: hmmm...now there's a thought | 23:03 |
| acoles | clayg: just about to push a new set of patches btw | 23:04 |
| clayg | acoles: very good - please do | 23:04 |
| *** siva_krish has quit IRC | 23:04 | |
| timburke | acoles: i'd be nervous about having it be a constant across *all* swift deployments...but across a *single* one, maybe... | 23:05 |
| clayg | patch 328207 looks right - no votes yet? | 23:05 |
| patchbot | clayg: https://review.openstack.org/#/c/328207/ - swift (feature/crypto-review) - Allow middleware to override metadata header checking | 23:05 |
| notmyname | timburke: did I miss something acoles said about the iv? | 23:05 |
| notmyname | acoles: after you push these, you're out until monday, right? | 23:05 |
| acoles | notmyname: I just published some comments on patch 328208 | 23:06 |
| patchbot | acoles: https://review.openstack.org/#/c/328208/ - swift (feature/crypto-review) - Enable object body and metadata encryption | 23:06 |
| acoles | clayg: yeah we have a review activity dip in the middle of the chain | 23:06 |
| timburke | notmyname: in his most-recent comments on...yeah, that one. "If we switch back to using object key whenever we use a derived key then I think there is no longer a requirement for IV's to be distinct w.r.t. other object paths" | 23:06 |
| acoles | some start at the top, some start at the bottom, only the dedicated make it to the middle and beyond ;) | 23:07 |
| notmyname | I'm in the middle now. not sure if I'll make it out alive | 23:07 |
| acoles | timburke: notmyname ... but I think they will be anyway by virtue of the padding | 23:08 |
| timburke | also, patch 328208 is just *big*. i keep coming back to it, biting off chunks, chewing on them a while... | 23:08 |
| patchbot | timburke: https://review.openstack.org/#/c/328208/ - swift (feature/crypto-review) - Enable object body and metadata encryption | 23:08 |
| timburke | i'm somewhere around the middle now | 23:08 |
| acoles | hmmm, except if path length is exactly 16 | 23:08 |
| timburke | acoles: but as you said, if we're using different keys, it shouldn't matter | 23:09 |
| acoles | teah | 23:09 |
| acoles | yeah | 23:09 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Support for http footers - Replication and EC https://review.openstack.org/328204 | 23:11 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Make container update override headers persistent https://review.openstack.org/328205 | 23:11 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Enable middleware to set metadata on object POST https://review.openstack.org/328206 | 23:11 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Allow middleware to override metadata header checking https://review.openstack.org/328207 | 23:11 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Enable object body and metadata encryption https://review.openstack.org/328208 | 23:11 |
| timburke | patches galore! | 23:12 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Add encryption overview doc https://review.openstack.org/328209 | 23:12 |
| acoles | timburke: but are they the right ones? makes me so nervous doing this | 23:12 |
| acoles | ok, known tasks still outstanding - rework the iv offset calculation ( jrichli is working on that ), torgomatic's unified Putter patch (I'm pinning its fate to the outcome of the uk referendum - to unify or not ;) | 23:14 |
| acoles | notmyname: timburke tdasilva kota_ jrichli ^^ | 23:14 |
| notmyname | thanks! | 23:15 |
| acoles | remind me of any other tasks | 23:15 |
| *** kei_yama has joined #openstack-swift | 23:15 | |
| *** rcernin has joined #openstack-swift | 23:16 | |
| notmyname | I don't know of any. and I think torgomatic's unified Putters is nice, but perhaps not a blocker | 23:16 |
| notmyname | acoles: you're offline until monday now, right? | 23:18 |
| torgomatic | yeah, if that had to wait until after the merge, that'd probably be okay | 23:18 |
| acoles | torgomatic: did you spot anything broken when you wrote that patch? | 23:23 |
| torgomatic | acoles: nothing broken, no | 23:23 |
| acoles | notmyname: I am around tomorrow but have an early finish (taking son to visit possible future university over weekend). Then back on Monday. | 23:23 |
| acoles | notmyname: that is, unless we brexit in which case the internet may be severed at the coast | 23:24 |
| notmyname | heh | 23:24 |
| acoles | notmyname: I'll be on the phone conf tomorrow | 23:25 |
| notmyname | ok | 23:25 |
| notmyname | right. I need to make sure my alarm is set | 23:25 |
| *** arch-nemesis has quit IRC | 23:26 | |
| *** foexle has quit IRC | 23:26 | |
| acoles | torgomatic: I did see some potential benefits in a single class while I was ironing out the conn resp references. Tactically it might save a load or rebase noise up the chain if we kept it for follow up. | 23:27 |
| acoles | s/or/of/ | 23:27 |
| acoles | good night | 23:29 |
| *** sheel has joined #openstack-swift | 23:29 | |
| *** acoles is now known as acoles_ | 23:30 | |
| *** McMurlock1 has quit IRC | 23:33 | |
| clayg | lol @ a number of .data files thinking they're something -> /srv/node4/sdb4/objects/890/fe9/dea1f52fbcc16540744db6850e4b8fe9/1466724125.69735.data: Sendmail frozen configuration - version m"\343\233q5\276C$\007\266\347:\341\015uFB\240\373\323\227D\273y\233T\316\341J\367A\262\033H\356\201\264\212\323I\354\331\3228\207\321fO/c\210U\006?Pg1,ml\201S | 23:36 |
| clayg | ^ from `file /srv/node4/sdb4/objects/890/fe9/dea1f52fbcc16540744db6850e4b8fe9/1466724125.69735.data` | 23:36 |
| *** ouchkernel has joined #openstack-swift | 23:38 | |
| *** adu has joined #openstack-swift | 23:40 | |
| *** dmorita has quit IRC | 23:41 | |
| *** dmorita has joined #openstack-swift | 23:45 | |
| *** chsc has quit IRC | 23:47 | |
| *** daemontool has joined #openstack-swift | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!