| *** DisneyRicky has quit IRC | 00:07 | |
| *** pandemicsyn2 has quit IRC | 00:07 | |
| *** esmute has quit IRC | 00:07 | |
| *** CybergeekDK has quit IRC | 00:07 | |
| *** shakamunyi has quit IRC | 00:07 | |
| *** pberis has joined #openstack-swift | 00:07 | |
| *** nottrobin_ has joined #openstack-swift | 00:07 | |
| *** zhiyan_ has joined #openstack-swift | 00:07 | |
| *** zacksh has joined #openstack-swift | 00:07 | |
| *** jdag___ has joined #openstack-swift | 00:07 | |
| *** openstack has joined #openstack-swift | 00:09 | |
| *** zhiyan_ is now known as zhiyan | 00:09 | |
| *** jdag___ is now known as jdag__ | 00:10 | |
| *** nottrobin_ is now known as nottrobin | 00:10 | |
| *** swat30 has quit IRC | 00:11 | |
| *** pberis1 has quit IRC | 00:11 | |
| *** mitz_ has quit IRC | 00:11 | |
| *** swat30_ is now known as swat30 | 00:11 | |
| *** dmorita has joined #openstack-swift | 00:33 | |
| *** shakamunyi has joined #openstack-swift | 00:54 | |
| *** early has quit IRC | 00:57 | |
| *** echevemaster has joined #openstack-swift | 00:58 | |
| *** echevemaster has joined #openstack-swift | 00:58 | |
| *** shakamunyi has quit IRC | 00:59 | |
| *** early has joined #openstack-swift | 01:02 | |
| *** addnull has joined #openstack-swift | 01:04 | |
| *** early has quit IRC | 01:05 | |
| *** early has joined #openstack-swift | 01:07 | |
| *** tsg has joined #openstack-swift | 01:11 | |
| *** early has quit IRC | 01:17 | |
| *** early has joined #openstack-swift | 01:21 | |
| *** tsg has quit IRC | 01:36 | |
| *** nosnos has joined #openstack-swift | 01:42 | |
| *** kota_ has joined #openstack-swift | 01:49 | |
| *** shakamunyi has joined #openstack-swift | 01:55 | |
| developer | still don't see any new updates in proxy.log | 01:58 |
|---|---|---|
| *** shakamunyi has quit IRC | 01:59 | |
| *** addnull has quit IRC | 02:12 | |
| *** addnull has joined #openstack-swift | 02:13 | |
| *** addnull has quit IRC | 02:17 | |
| *** addnull has joined #openstack-swift | 02:18 | |
| *** sungju has quit IRC | 02:36 | |
| developer | <peluse> Here is the error response I get back when I try to do the HEAD request | 02:53 |
| developer | +peluse: https://gist.github.com/anonymous/4ff3fda4c028aac1eb48 | 02:54 |
| *** shakamunyi has joined #openstack-swift | 02:56 | |
| developer | +peluse: Also left some other updates while you were away. | 02:59 |
| developer | (scroll above) | 02:59 |
| developer | (to see) | 02:59 |
| *** shakamunyi has quit IRC | 03:00 | |
| *** addnull has quit IRC | 03:06 | |
| *** addnull has joined #openstack-swift | 03:06 | |
| *** addnull has quit IRC | 03:10 | |
| *** addnull has joined #openstack-swift | 03:21 | |
| *** addnull has quit IRC | 03:24 | |
| *** addnull has joined #openstack-swift | 03:24 | |
| *** tsg has joined #openstack-swift | 03:50 | |
| *** shakamunyi has joined #openstack-swift | 03:57 | |
| *** shakamunyi has quit IRC | 04:01 | |
| *** addnull has quit IRC | 04:05 | |
| *** addnull has joined #openstack-swift | 04:05 | |
| *** fifieldt_ has quit IRC | 04:08 | |
| *** addnull has quit IRC | 04:10 | |
| *** tumf has joined #openstack-swift | 04:13 | |
| *** ppai has joined #openstack-swift | 04:15 | |
| *** chandankumar has joined #openstack-swift | 04:32 | |
| *** chandankumar has quit IRC | 04:53 | |
| *** shakamunyi has joined #openstack-swift | 04:58 | |
| *** tumf has quit IRC | 04:59 | |
| *** tumf has joined #openstack-swift | 05:00 | |
| *** chandankumar has joined #openstack-swift | 05:01 | |
| *** shakamunyi has quit IRC | 05:02 | |
| *** addnull has joined #openstack-swift | 05:03 | |
| *** tumf_ has joined #openstack-swift | 05:04 | |
| *** tumf_ has quit IRC | 05:05 | |
| *** tumf_ has joined #openstack-swift | 05:06 | |
| *** tumf has quit IRC | 05:07 | |
| *** chandankumar has quit IRC | 05:09 | |
| *** kopparam has joined #openstack-swift | 05:13 | |
| *** chandankumar has joined #openstack-swift | 05:14 | |
| *** tumf_ has quit IRC | 05:38 | |
| *** chandankumar has quit IRC | 05:40 | |
| *** tumf has joined #openstack-swift | 05:44 | |
| *** shakamunyi has joined #openstack-swift | 05:58 | |
| *** tumf has quit IRC | 06:01 | |
| *** shakamunyi has quit IRC | 06:03 | |
| *** k4n0 has joined #openstack-swift | 06:04 | |
| *** sungju has joined #openstack-swift | 06:05 | |
| *** echevemaster has quit IRC | 06:12 | |
| *** Anju has joined #openstack-swift | 06:16 | |
| *** joeljwright has joined #openstack-swift | 06:17 | |
| *** chandankumar has joined #openstack-swift | 06:19 | |
| openstackgerrit | Yuan Zhou proposed a change to openstack/swift: Use mkstemp() instead of deprecated mktemp() https://review.openstack.org/118084 | 06:23 |
| *** slo_ has quit IRC | 06:26 | |
| *** haomai___ has quit IRC | 06:35 | |
| *** chandankumar has quit IRC | 06:38 | |
| *** chandankumar has joined #openstack-swift | 06:43 | |
| *** haomaiwang has joined #openstack-swift | 06:54 | |
| *** chandankumar has quit IRC | 06:54 | |
| *** chandankumar has joined #openstack-swift | 06:54 | |
| *** Trixboxer has joined #openstack-swift | 06:55 | |
| *** shakamunyi has joined #openstack-swift | 06:59 | |
| *** kopparam has quit IRC | 07:04 | |
| *** shakamunyi has quit IRC | 07:04 | |
| *** kopparam has joined #openstack-swift | 07:04 | |
| *** kopparam has quit IRC | 07:09 | |
| *** sungju has quit IRC | 07:27 | |
| *** ksuni has joined #openstack-swift | 07:27 | |
| *** tab___ has joined #openstack-swift | 07:27 | |
| *** ksuni has quit IRC | 07:28 | |
| *** chandankumar has quit IRC | 07:28 | |
| *** kopparam has joined #openstack-swift | 07:33 | |
| *** chandankumar has joined #openstack-swift | 07:41 | |
| *** nshaikh has joined #openstack-swift | 07:52 | |
| *** shakamunyi has joined #openstack-swift | 08:00 | |
| *** geaaru has joined #openstack-swift | 08:01 | |
| *** tsg has quit IRC | 08:03 | |
| *** kota_ has quit IRC | 08:04 | |
| *** shakamunyi has quit IRC | 08:05 | |
| *** sungju has joined #openstack-swift | 08:18 | |
| *** Dafna has joined #openstack-swift | 08:28 | |
| *** sungju has quit IRC | 08:29 | |
| *** kopparam has quit IRC | 08:30 | |
| *** kopparam has joined #openstack-swift | 08:31 | |
| *** Midnightmyth has joined #openstack-swift | 08:34 | |
| *** nshaikh has quit IRC | 08:47 | |
| *** k4n0 has quit IRC | 08:47 | |
| *** aix has joined #openstack-swift | 08:56 | |
| *** shakamunyi has joined #openstack-swift | 09:01 | |
| *** chandankumar has quit IRC | 09:02 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/swift: Merge "Allow empty reseller prefixes in domain_remap" https://review.openstack.org/118108 | 09:04 |
| *** k4n0 has joined #openstack-swift | 09:05 | |
| *** shakamunyi has quit IRC | 09:05 | |
| *** mkollaro has joined #openstack-swift | 09:20 | |
| *** k4n0 has quit IRC | 10:00 | |
| *** shakamunyi has joined #openstack-swift | 10:02 | |
| *** shakamunyi has quit IRC | 10:07 | |
| *** kopparam_ has joined #openstack-swift | 10:11 | |
| *** kopparam has quit IRC | 10:13 | |
| *** k4n0 has joined #openstack-swift | 10:13 | |
| *** kopparam_ has quit IRC | 10:17 | |
| *** kopparam has joined #openstack-swift | 10:18 | |
| *** kopparam has quit IRC | 10:23 | |
| ctennis | developer: the line "Invalid path: /node_for_swift/sdf1/807/AUTH_system" is interesting..do you have the drive mounted at /node_for_swift/sdf1 ? | 10:33 |
| ctennis | the corresponding log file showing what the swift processes were doing would be more interesting I think..what error they're reporting behind the scenes that causes you to get a 400 back | 10:33 |
| *** developer has quit IRC | 10:37 | |
| *** dmorita has quit IRC | 10:57 | |
| *** shakamunyi has joined #openstack-swift | 11:02 | |
| *** shakamunyi has quit IRC | 11:07 | |
| *** kopparam has joined #openstack-swift | 11:14 | |
| *** addnull has quit IRC | 11:23 | |
| *** chandankumar has joined #openstack-swift | 11:26 | |
| *** kopparam has quit IRC | 11:28 | |
| *** kopparam has joined #openstack-swift | 11:29 | |
| *** zul has joined #openstack-swift | 11:36 | |
| *** chandankumar has quit IRC | 11:39 | |
| *** addnull has joined #openstack-swift | 11:45 | |
| *** addnull has quit IRC | 11:55 | |
| *** shakamunyi has joined #openstack-swift | 12:03 | |
| *** marcusvrn has joined #openstack-swift | 12:06 | |
| *** shakamunyi has quit IRC | 12:08 | |
| *** Midnightmyth has quit IRC | 12:10 | |
| *** kopparam has quit IRC | 12:22 | |
| *** kopparam has joined #openstack-swift | 12:23 | |
| *** kopparam has quit IRC | 12:27 | |
| *** nosnos has quit IRC | 12:32 | |
| *** nosnos has joined #openstack-swift | 12:33 | |
| *** ppai has quit IRC | 12:36 | |
| *** nosnos has quit IRC | 12:37 | |
| *** addnull has joined #openstack-swift | 12:38 | |
| *** addnull has quit IRC | 12:46 | |
| *** ppai has joined #openstack-swift | 12:50 | |
| *** bkopilov has joined #openstack-swift | 13:04 | |
| *** shakamunyi has joined #openstack-swift | 13:04 | |
| *** shakamunyi has quit IRC | 13:09 | |
| peluse | developer: I'll be on and off today throughout the day, let me know when you're around. yeah, the invalid path thing looks like it could be what's behind your error | 13:17 |
| *** silverpuff has quit IRC | 13:39 | |
| peluse | developer: and also I didn't see the path /node_for_swift specified anyone in your .conf files either. if you deviated from the directions in the link you sent in step 2 of "Configure the Storage nodes" then you could have some problems because the default path for devices is /srv/node if not specified in the .conf files and it looks like you did somthing else | 13:40 |
| peluse | develoer: I would suggest instead of trying to go configure things for the path you chose, just go follow step2 exactly and then restart your services and try again | 13:41 |
| peluse | developer: note that the devices path is also specified in /etc/rsyncd.conf: | 13:42 |
| *** gindating has joined #openstack-swift | 13:44 | |
| *** bvandenh has joined #openstack-swift | 13:47 | |
| ctennis | peluse: he's gone from IRC unfortunately | 13:48 |
| *** openstack has joined #openstack-swift | 13:57 | |
| *** openstackstatus has joined #openstack-swift | 13:58 | |
| *** ChanServ sets mode: +v openstackstatus | 13:58 | |
| *** Anju has quit IRC | 13:59 | |
| peluse | ctennis: yeah, I figure maybe he's using a bouner or something and will see it later :) | 14:00 |
| *** zaitcev has joined #openstack-swift | 14:03 | |
| *** ChanServ sets mode: +v zaitcev | 14:03 | |
| *** shakamunyi has joined #openstack-swift | 14:05 | |
| *** shakamunyi has quit IRC | 14:09 | |
| *** Alex_Gaynor_ is now known as Alex_Gaynor | 14:18 | |
| *** Alex_Gaynor has quit IRC | 14:18 | |
| *** Alex_Gaynor has joined #openstack-swift | 14:18 | |
| *** aix has quit IRC | 14:36 | |
| *** k4n0 has quit IRC | 14:37 | |
| *** HenryG has joined #openstack-swift | 14:57 | |
| *** aix has joined #openstack-swift | 15:01 | |
| *** shakamunyi has joined #openstack-swift | 15:06 | |
| *** shakamunyi has quit IRC | 15:10 | |
| *** ppai has quit IRC | 15:13 | |
| *** ChanServ changes topic to "Swift 2.1.0 https://launchpad.net/swift/+milestone/2.1.0 | Swift Review Dashboard: http://bit.ly/1iVBigF | Priority Reviews: https://wiki.openstack.org/wiki/Swift/PriorityReviews" | 15:17 | |
| *** bvandenh has quit IRC | 15:29 | |
| *** bkopilov has quit IRC | 15:47 | |
| openstackgerrit | anju Tiwari proposed a change to openstack/swift: Added a check for limit value https://review.openstack.org/118186 | 15:57 |
| *** shakamunyi has joined #openstack-swift | 16:06 | |
| *** shakamunyi has quit IRC | 16:11 | |
| *** aix has quit IRC | 16:23 | |
| goodes | peluse: good morning | 16:27 |
| *** aix has joined #openstack-swift | 16:36 | |
| *** mkollaro has quit IRC | 16:41 | |
| *** Anju has joined #openstack-swift | 16:56 | |
| Anju | hi notmyname | 16:57 |
| Anju | I see your comments ..I did not get the metacoment part | 16:58 |
| *** shakamunyi has joined #openstack-swift | 17:07 | |
| notmyname | Anju: ah, just saying that I'm not sure what the end-user/client impact will be | 17:07 |
| notmyname | Anju: that is, the current behavior is to ignore malformed limit queries. your new behavior is to return an error | 17:07 |
| notmyname | Anju: I'd love your comments on that, but I also want to get the feedback of other core reviewers too | 17:08 |
| Anju | notmyname: at least postive | 17:08 |
| Anju | :) | 17:08 |
| notmyname | Anju: but note that my comment on that in no way reflects the code. just curious about any migration path | 17:08 |
| notmyname | Anju: one thing you could do that would be helpful is to look a swiftclient and see how or if it handles that case. | 17:09 |
| notmyname | and leave a comment about it in your patch | 17:09 |
| notmyname | Anju: what does python-swiftclient do if a non-int is given for the limit parameter | 17:09 |
| Anju | notmyname: actually I got this error because I want to add a limit options there also | 17:10 |
| notmyname | :-) | 17:10 |
| Anju | I was getting this error | 17:10 |
| Anju | so I checked swift code | 17:10 |
| Anju | then caught thsi | 17:10 |
| Anju | this | 17:10 |
| Anju | finding a bug in swift is so hard | 17:11 |
| Anju | :) | 17:11 |
| notmyname | heh. don't worry. they are there :-) | 17:11 |
| Anju | :) | 17:11 |
| notmyname | ...all my code hasn't been rewritten out of swift yet ;-) | 17:12 |
| Anju | :) :) | 17:12 |
| *** shakamunyi has quit IRC | 17:12 | |
| Anju | I will change the code part .. If..else..If you want | 17:12 |
| Anju | but I hope cores accept this | 17:13 |
| Anju | fingures crossed | 17:13 |
| ahale | quick question on proposing change to swift .. does it have to have a bug/bp ? or can i just drop a patch ? | 17:16 |
| notmyname | sure. actually, I like it, in general. I just want to make sure that we don't inadvertently break the ecosystem | 17:16 |
| notmyname | ahale: just drop a patch :-) | 17:16 |
| ahale | cool :) | 17:17 |
| notmyname | ahale: bugs and blueprints are nice to have. and I'll probably have to go create it after the fact if you don't. but don't consider that a barrier to entry (IMO) | 17:17 |
| ahale | the contributing to openstack and gerrit workflow pages are a long hard read.. | 17:17 |
| notmyname | ahale: normally, I'll only ask for bugs/blueprints if it's something more significant, especially something that needs to be tracked external to swift | 17:17 |
| ahale | ahh nah its just a saio config change | 17:18 |
| ahale | nothing significant :) | 17:18 |
| notmyname | ahale: oh. well then certainly don't fight with LP to get that in :-) | 17:18 |
| Anju | Agree notmyname | 17:18 |
| notmyname | Anju: if you would have had the test I would have left it as +0 until more people could weigh in (note that it's a US holiday today, so it will be at least tomorrow before people are likely to look at it) | 17:19 |
| ahale | i figured it makes sense to have bind_ip set, since the ring-builder and curl lines use it anyway so no impact on tests or docs, and it saves putting a public + default password box up if you're doing it somewhere with public ips | 17:20 |
| ahale | *set to localhost | 17:20 |
| notmyname | cool | 17:20 |
| notmyname | ahale: one thing we want to do is make the bind_port required (no default, must be set in configs) so that we can later change the default port in configs | 17:21 |
| notmyname | ahale: turns out that some people get really upset* that swift's default ports conflict with x.org | 17:22 |
| ahale | hehe, finally get off the port 6000 :) | 17:22 |
| notmyname | *well, annoyed | 17:22 |
| notmyname | ahale: I'm assuming you have all the ports explicitly set anyway already :-) | 17:22 |
| ahale | they're all set in the saio at least, for the crazy ports | 17:23 |
| Anju | yes notmyname ..I will..No body will giev a single eye with your -1 :-) | 17:25 |
| notmyname | ahale: don't forget the sample saio configs in doc/saio if you are changing something there | 17:25 |
| ahale | yeah thats the ones I am changing! are there others that should be more obvious notmyname ? :) | 17:26 |
| *** mahatic has joined #openstack-swift | 17:26 | |
| notmyname | ahale: no IIRC all the docs directly import those now. maybe there was something in the surrounding text | 17:26 |
| ahale | ahh | 17:26 |
| ahale | yeah no the docs do import them so its just that - dont think it even needs a mention in the text | 17:27 |
| *** acoles is now known as acoles_away | 17:28 | |
| Anju | notmyname: sorry I am asking you more,,are you not happy with condition ? you said to change the if ..else part also ? | 17:33 |
| notmyname | Anju: oh. that's less important. I prefer it with the else, but I should have added a "nit" tag to those comments. | 17:34 |
| notmyname | Anju: so yeah I think it would be nice, and I think it's clearer with the else. but you're the one writing it :-) | 17:35 |
| *** mahatic has quit IRC | 17:36 | |
| notmyname | hmm..oops. teh proxy sample config says the default is port 80. the default is actually 8080 | 17:36 |
| Anju | I was happy with that...but will see :P | 17:37 |
| ahale | ugh gerrit docs are very confusing - they suggest Change-Id is meant to be added by a post hook, and if its missing directs you to a part of docs that has no mention of it >.< | 17:42 |
| notmyname | ahale: did you install git-review? | 17:42 |
| ahale | yeah, it added a remote fine as well | 17:42 |
| notmyname | ahale: yeah, when you do the initial `git review -s` it sets up the post commit hook to add the change-id to the commit message | 17:43 |
| ahale | lemme try again from scratch | 17:43 |
| *** mahatic has joined #openstack-swift | 17:49 | |
| *** geaaru has quit IRC | 17:59 | |
| *** tsg has joined #openstack-swift | 18:02 | |
| ahale | oh i am an idiot | 18:03 |
| openstackgerrit | Andrew Hale proposed a change to openstack/swift: Only bind SAIO daemons to localhost https://review.openstack.org/118197 | 18:04 |
| ahale | ew, can you easily change a commit msg ? like, adding linebreaks | 18:05 |
| Anju | again checkout the samwe..and then git commit -a --amend | 18:07 |
| *** shakamunyi has joined #openstack-swift | 18:08 | |
| notmyname | -a is only needed of you're adding new files to the commit | 18:09 |
| ahale | so just a git commit --amend ? | 18:10 |
| notmyname | ahale: yes | 18:10 |
| notmyname | ahale: or I can do it in gerrit | 18:10 |
| notmyname | ahale: whichever you prefer | 18:10 |
| ahale | oh if you dont mind that would be awesome | 18:10 |
| notmyname | kk | 18:10 |
| openstackgerrit | John Dickinson proposed a change to openstack/swift: Only bind SAIO daemons to localhost https://review.openstack.org/118197 | 18:11 |
| ahale | this kind of git use is terrying to ops like me | 18:11 |
| ahale | thanks! | 18:11 |
| notmyname | :-) | 18:11 |
| notmyname | there's a little "edit the commit message in the gerrit UI. that's the first time I've used it | 18:11 |
| ahale | oh nice :) | 18:12 |
| *** shakamunyi has quit IRC | 18:13 | |
| *** Anju has quit IRC | 18:14 | |
| *** mahatic_ has joined #openstack-swift | 18:19 | |
| *** mahatic has quit IRC | 18:20 | |
| ahale | notmyname: looks like the swiftstack ci thingy is having trouble with pypi | 18:20 |
| notmyname | ahale: ya, I just asked charz about it | 18:20 |
| ahale | :) | 18:21 |
| notmyname | but it's really late/early for him right now :-) | 18:21 |
| notmyname | (taipei) | 18:22 |
| openstackgerrit | John Dickinson proposed a change to openstack/swift: make the bind_port config setting required https://review.openstack.org/118200 | 18:25 |
| notmyname | there's a first attempt at making the bind_port required | 18:25 |
| *** aix has quit IRC | 18:36 | |
| peluse | notmyname: just installed a fresh SAIO on trusty and seeing 5 errors in unittests complaining about 'long int too large to congert to int' and wondering if you've heard of this... | 18:37 |
| ahale | fwiw, I did a fresh one on precise yesterday and it was fine | 18:39 |
| peluse | thanks, no issues w/precise. wondering if there's thing here with python 2.7.3 (precise) vs 2.7.6 (what you get w/trusty) or if I just shouldn't be working on a holiday | 18:40 |
| notmyname | peluse: is this one of your old 16 bit machines? ;-) | 18:40 |
| notmyname | peluse: IOW, no idea :-) | 18:41 |
| peluse | notmyname, heh. OK, I'll maybe look at it closer later after a beer or something :) | 18:41 |
| notmyname | peluse: but the biggest config values I can think of would be the constraints set in swift.conf. check that you didn't add another digit to max_object_size or something | 18:41 |
| notmyname | peluse: maybe we cast that to int "because sys.maxint is enough for anybody" | 18:42 |
| notmyname | (maxint is 9223372036854775807 on my saio) | 18:43 |
| notmyname | both 64bit | 18:43 |
| notmyname | both == mac and saio | 18:44 |
| peluse | wow man, for my trusty is 2147483647 and for my precise its same as yours | 18:44 |
| peluse | WTF happened there! | 18:44 |
| notmyname | if you've got a 32 bit vm, it will be 4294967295 (5gb == 5368709120) | 18:45 |
| notmyname | singed ints on 32 bit | 18:45 |
| notmyname | 2**31-1 | 18:45 |
| peluse | ah hell, I grabbed the wrong ISO. thanks! | 18:46 |
| peluse | &*(&(*!HKJH&^!!!! | 18:46 |
| notmyname | that's an interesting failure mode, though. we explicity cast to int() in many places with the general expectation that header values (eg content-length) and constraints will be representable there | 18:47 |
| notmyname | not exactly 10-year future proof, I guess | 18:47 |
| peluse | yup, the error was not obvious, I took a quick look at the code. Will reisntall my new insteall, argh | 18:48 |
| * notmyname would love a network connection that can transfer more than 2**64 bits in one stream | 18:48 | |
| Alex_Gaynor | int() will happilly return a long if you send a value too large | 18:48 |
| Alex_Gaynor | >>>> int("3413432480932497823147890214214213") | 18:48 |
| Alex_Gaynor | 3413432480932497823147890214214213L | 18:48 |
| notmyname | Alex_Gaynor: hmm...so I wonder where that is getting thrown in swift | 18:49 |
| peluse | master: File "/home/peluse/swift/test/unit/__init__.py", line 755, | 18:49 |
| peluse | but I've already blown it away so don't ask for more details :) | 18:50 |
| *** Jouva has joined #openstack-swift | 19:08 | |
| Jouva | So I'm fairly new to OpenStack but know just a few basics about it. Have not deployed anything yet. Not going for a huge deployment (like 1-3 TB hosted in 2 or 3 locations). For such a small deployment is it worth it? If so, what sort of hardware could one get away with for CPU/Mem for the storage and proxy machines? We don't expect heavy amounts of traffic. Nothing constant. But we | 19:11 |
| Jouva | want redundancy (physical storage and network wise) and the ability to grow our storage | 19:11 |
| zaitcev | I pass on suggesting on this. I don't deal with geographic distribution among "2 or 3 locations". | 19:13 |
| openstackgerrit | anju Tiwari proposed a change to openstack/swift: Added a check for limit value https://review.openstack.org/118186 | 19:15 |
| notmyname | Jouva: general guidelines are about 1GB of ram per spindle and about 1 core for every 2-3 spindles. it varies greatly depending on your use case though | 19:18 |
| Jouva | notmyname: Is that for the storage itself? | 19:19 |
| notmyname | Jouva: 1-3TB is generally really small for swift clusters (normally you'd start with a machine with 12-24 data drives. with 3x replication, that gives you more than your looking for) | 19:19 |
| Jouva | Got ya | 19:19 |
| notmyname | Jouva: yeah. large deployments could specialize in as many as 3 different SKUs, but small deployments can easily run everything on one box | 19:20 |
| notmyname | or one SKU. eg have 2 boxes for some HA, but they each run all the swift processes | 19:20 |
| notmyname | Jouva: so if you've got one box with 12 drives, a quad core cpu, 16gb ram, and a 10g nic, you'll have a nice all-in-one box | 19:22 |
| notmyname | Jouva: and with 12 4t drives, that will give you a total of 4tb usable storage with 3x replication | 19:23 |
| Jouva | Thought that'd give me more, but ok! | 19:24 |
| Jouva | I know there's redundancy involved and all that, but I thought 3x replication would cut the storage down by 1/3 not by 1/12th | 19:25 |
| notmyname | ah, my math is bad | 19:25 |
| Jouva | 12 1TBs would give me 4TB, right? | 19:25 |
| notmyname | yes | 19:25 |
| Jouva | Ok | 19:25 |
| notmyname | sorry. I'll blame that on the holiday...yeah..that's it...the holiday | 19:26 |
| zaitcev | IN AMERICA | 19:26 |
| Jouva | That excuse is Good Enough For Me(tm) | 19:26 |
| zaitcev | So, what about the 3 locations? Just deploy 3 clusters and have the app select, like Amazon S3 regions? | 19:27 |
| zaitcev | Well, 3 boxes | 19:27 |
| zaitcev | at this size | 19:27 |
| notmyname | Jouva: for base 10 (marketing) to base 2 (formatted) space, allowing for 80% full, it's pretty close to taking the total TB on the box and dividing by 4. eg 100 6TB hard drives is about 150TB usable) | 19:27 |
| Jouva | Right right | 19:27 |
| Jouva | I know it's not exact, but the first number you said wasn't right. | 19:28 |
| Jouva | But with 12 1TBs it's 4TB-ish | 19:28 |
| notmyname | zaitcev: depends on the use case | 19:28 |
| notmyname | for multiple regions, you can either allow the client to choose, or you can let swift manage the replication between them and expose it as a single endpoint | 19:28 |
| notmyname | basically, do you need locality of access (lower latency) or do you need some DR story in case one region goes offline | 19:29 |
| ahale | I always think looking at the max available storage is dangerous, since getting a full swift cluster is such a painful thing to deal with you want to knock 20% off when reporting to people who pay for it | 19:30 |
| notmyname | ahale: yes :-) | 19:31 |
| notmyname | ahale: something I've learned, though, is that a lot of the traditional storage options that swift replaces tend to start really hurting after 50-60% full (anecdotal) | 19:32 |
| ahale | oh really | 19:32 |
| notmyname | yeah. not to mention that some need nightly "defrags" that mean you don't want end-user requests to it | 19:34 |
| ahale | I've seen container dbs that could have benefitted from a scheduled vacuum :) | 19:35 |
| notmyname | :-) | 19:36 |
| ahale | oh, are there any plans for container/account storage policies ? | 19:36 |
| ahale | sometimes it could be cool to split out some containers onto different hardware, .expiring for example | 19:37 |
| notmyname | not right now | 19:37 |
| ahale | are blueprints the way to suggest features ? (not that one, another"it would be cool if" I just remembered) | 19:39 |
| ahale | they seem kinda formal | 19:39 |
| ahale | i guess a bug report that you can't do somrthing could work | 19:40 |
| notmyname | ahale: the specs documents are supposed to be a better way to do that. clone the repo, add/edit a text file, and submit through gerrit | 19:49 |
| notmyname | I've got a list of small niceties that I'd like to track somewhere. | 19:49 |
| notmyname | I don't know if wiki page, catch-all specs document, or blueprints is the best way | 19:49 |
| ahale | I'm not really sure | 19:50 |
| notmyname | I'll probably either do a wiki page or blueprints. or a wiki with blueprints linked... | 19:51 |
| notmyname | right now it's all in an evernote doc :-) | 19:51 |
| ahale | not being a dev, something that doesn't carry a feeling of "heres work I think someone should do" would be nice - or a way to gain concensus something is worth adding before adding it | 19:51 |
| ahale | updaing official docs straight out feels a little presumptous :) | 19:51 |
| notmyname | :-) | 19:51 |
| notmyname | I don't think the specs should necessarily be seen that way. in fact, they were first proposed at an ops summit to figure out how ops can have better feedback into the dev community | 19:53 |
| ahale | maybe it would feel different in a less established project | 19:54 |
| jokke_ | notmyname: gratz for another successful release ;) | 19:56 |
| ahale | also I'm probably in a funny position, anything related to my business would go through my devs normally, things from my daydreaming about features isn't really relevent for that route | 19:57 |
| *** ChanServ changes topic to "Swift 2.1.0 https://launchpad.net/swift/+milestone/2.1.0 | Swift Review Dashboard: http://bit.ly/1iVBigF | Priority Reviews: https://wiki.openstack.org/wiki/Swift/PriorityReviews | Ideas: https://wiki.openstack.org/wiki/Swift/ideas" | 20:34 | |
| notmyname | ahale: new link in the /topic | 20:34 |
| notmyname | ahale: https://wiki.openstack.org/wiki/Swift/ideas | 20:34 |
| ahale | aha cool stuff | 20:34 |
| ahale | oh label mounting in docs is very low hanging fruit | 20:35 |
| *** swills has quit IRC | 20:49 | |
| *** stenchlarge has joined #openstack-swift | 21:01 | |
| *** joeljwright has quit IRC | 21:02 | |
| *** swills has joined #openstack-swift | 21:11 | |
| *** tdasilva has quit IRC | 21:16 | |
| openstackgerrit | anju Tiwari proposed a change to openstack/swift: Added a check for limit value https://review.openstack.org/118186 | 21:28 |
| *** samwebster has quit IRC | 21:33 | |
| *** samwebster has joined #openstack-swift | 21:34 | |
| *** echevemaster has joined #openstack-swift | 21:45 | |
| *** haomaiwang has quit IRC | 22:11 | |
| *** haomaiwa_ has joined #openstack-swift | 22:30 | |
| *** haomaiwa_ has quit IRC | 23:01 | |
| *** haomaiwang has joined #openstack-swift | 23:02 | |
| mattoliverau | Morning all | 23:12 |
| *** HenryG has quit IRC | 23:13 | |
| *** haomaiwang has quit IRC | 23:22 | |
| *** haomaiwa_ has joined #openstack-swift | 23:25 | |
| *** tsg has quit IRC | 23:26 | |
| *** HenryG has joined #openstack-swift | 23:28 | |
| *** fifieldt has joined #openstack-swift | 23:32 | |
| *** haomaiwa_ has quit IRC | 23:34 | |
| *** haomaiwa_ has joined #openstack-swift | 23:35 | |
| *** haomaiwa_ has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!