| *** lcurtis has quit IRC | 00:04 | |
| *** gyee has quit IRC | 00:05 | |
| *** annegent_ has joined #openstack-swift | 00:22 | |
| *** annegent_ has quit IRC | 00:28 | |
| *** ho has joined #openstack-swift | 00:28 | |
| *** annegent_ has joined #openstack-swift | 00:31 | |
| *** oomichi has joined #openstack-swift | 00:32 | |
| *** echevemaster has quit IRC | 00:35 | |
| *** annegent_ has quit IRC | 00:36 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift: Only move too-close-together replicas when they can spread out. https://review.openstack.org/140879 | 00:37 |
|---|---|---|
| torgomatic | clayg: ^^ | 00:37 |
| torgomatic | swifterdarrell: ctennis: ^^ | 00:39 |
| *** dmorita has joined #openstack-swift | 00:39 | |
| *** Masahiro has joined #openstack-swift | 00:40 | |
| *** Masahiro has quit IRC | 00:45 | |
| *** miurahr has joined #openstack-swift | 00:47 | |
| *** echevemaster has joined #openstack-swift | 00:50 | |
| ctennis | torgomatic: I feel like I gained some type of object storage certification just having read the analysis | 00:54 |
| torgomatic | ctennis: I certainly feel certifiable after figuring that one out | 00:55 |
| *** miurahr has quit IRC | 01:03 | |
| *** miurahr has joined #openstack-swift | 01:06 | |
| *** ho__ has joined #openstack-swift | 01:19 | |
| *** miurahr has quit IRC | 01:21 | |
| *** miurahr has joined #openstack-swift | 01:22 | |
| *** Masahiro has joined #openstack-swift | 01:24 | |
| *** kopparam has joined #openstack-swift | 01:25 | |
| *** rmcall has joined #openstack-swift | 01:26 | |
| *** ahonda has quit IRC | 01:29 | |
| *** miurahr has quit IRC | 01:29 | |
| *** ahonda has joined #openstack-swift | 01:30 | |
| *** annegent_ has joined #openstack-swift | 01:36 | |
| *** annegent_ has quit IRC | 01:41 | |
| *** rmcall has quit IRC | 01:42 | |
| *** ho has quit IRC | 01:45 | |
| *** ho__ has quit IRC | 01:45 | |
| *** ho has joined #openstack-swift | 01:45 | |
| *** rmcall has joined #openstack-swift | 01:46 | |
| *** rmcall has quit IRC | 01:47 | |
| *** rmcall has joined #openstack-swift | 01:51 | |
| *** annegent_ has joined #openstack-swift | 01:52 | |
| *** kopparam has quit IRC | 01:57 | |
| *** tellesnobrega_ has joined #openstack-swift | 01:59 | |
| *** haomaiwang has joined #openstack-swift | 02:01 | |
| *** tellesnobrega_ has quit IRC | 02:17 | |
| *** david-lyle is now known as david-lyle_afk | 02:30 | |
| *** annegent_ has quit IRC | 02:31 | |
| *** shakamunyi_ has joined #openstack-swift | 02:37 | |
| *** Masahiro has quit IRC | 02:42 | |
| *** rmcall has quit IRC | 02:45 | |
| *** shakamunyi_ has quit IRC | 02:48 | |
| *** Masahiro has joined #openstack-swift | 02:49 | |
| *** tellesnobrega_ has joined #openstack-swift | 02:58 | |
| *** shakamunyi_ has joined #openstack-swift | 03:02 | |
| *** Masahiro has quit IRC | 03:08 | |
| *** shakamunyi_ has quit IRC | 03:12 | |
| *** Masahiro has joined #openstack-swift | 03:16 | |
| *** kopparam has joined #openstack-swift | 03:42 | |
| *** jyoti-ranjan has joined #openstack-swift | 03:46 | |
| *** tellesnobrega_ has quit IRC | 03:49 | |
| *** X019 has quit IRC | 04:08 | |
| *** Masahiro has quit IRC | 04:09 | |
| *** silor has joined #openstack-swift | 04:10 | |
| *** Masahiro has joined #openstack-swift | 04:12 | |
| *** kopparam has quit IRC | 04:15 | |
| *** silor has quit IRC | 04:17 | |
| *** silor has joined #openstack-swift | 04:18 | |
| *** annegent_ has joined #openstack-swift | 04:24 | |
| *** Masahiro has quit IRC | 04:25 | |
| *** annegent_ has quit IRC | 04:25 | |
| *** Masahiro has joined #openstack-swift | 04:33 | |
| *** echevemaster has quit IRC | 04:45 | |
| *** echevemaster has joined #openstack-swift | 04:55 | |
| *** addnull has joined #openstack-swift | 05:13 | |
| *** ppai has joined #openstack-swift | 05:17 | |
| *** rmcall has joined #openstack-swift | 05:33 | |
| *** jyoti-ranjan has quit IRC | 05:43 | |
| *** echevemaster has quit IRC | 05:57 | |
| *** kopparam has joined #openstack-swift | 06:00 | |
| *** Masahiro has quit IRC | 06:00 | |
| *** Masahiro has joined #openstack-swift | 06:03 | |
| mattoliverau | I'm calling it a day you quiet old swift land you.. have a great night! | 06:06 |
| *** nshaikh has joined #openstack-swift | 06:14 | |
| *** Masahiro has quit IRC | 06:34 | |
| *** rmcall has quit IRC | 06:43 | |
| *** SkyRocknRoll has joined #openstack-swift | 07:31 | |
| *** zaitcev has quit IRC | 07:31 | |
| *** jyoti-ranjan has joined #openstack-swift | 07:37 | |
| *** k4n0 has joined #openstack-swift | 07:44 | |
| *** Masahiro has joined #openstack-swift | 07:44 | |
| *** zul has quit IRC | 07:49 | |
| *** Masahiro has quit IRC | 07:51 | |
| *** bsdkurt has quit IRC | 07:52 | |
| *** bsdkurt has joined #openstack-swift | 07:54 | |
| *** Masahiro has joined #openstack-swift | 07:56 | |
| *** nellysmitt has joined #openstack-swift | 08:00 | |
| *** zul has joined #openstack-swift | 08:03 | |
| *** rledisez has joined #openstack-swift | 08:05 | |
| openstackgerrit | Keshava Bharadwaj proposed openstack/swift: Adds console logging to swift-drive-audit https://review.openstack.org/140972 | 08:19 |
| *** nellysmitt has quit IRC | 08:20 | |
| *** haomaiwang has quit IRC | 08:21 | |
| *** haomaiwang has joined #openstack-swift | 08:21 | |
| *** addnull has quit IRC | 08:29 | |
| *** mkerrin has quit IRC | 08:29 | |
| *** haomaiwang has quit IRC | 08:31 | |
| *** haomaiwang has joined #openstack-swift | 08:33 | |
| *** acoles_away is now known as acoles | 08:40 | |
| *** geaaru has joined #openstack-swift | 08:54 | |
| *** jordanP has joined #openstack-swift | 08:57 | |
| *** kopparam has quit IRC | 09:01 | |
| *** jistr has joined #openstack-swift | 09:06 | |
| *** nellysmi_ has joined #openstack-swift | 09:23 | |
| *** mkerrin has joined #openstack-swift | 09:27 | |
| *** kopparam has joined #openstack-swift | 09:29 | |
| *** Masahiro has quit IRC | 09:29 | |
| *** Masahiro has joined #openstack-swift | 09:30 | |
| *** ppai has quit IRC | 09:39 | |
| *** silor has quit IRC | 09:42 | |
| *** silor has joined #openstack-swift | 09:43 | |
| *** Masahiro has quit IRC | 09:52 | |
| *** ppai has joined #openstack-swift | 09:53 | |
| *** nellysmi_ is now known as nellysmitt | 10:02 | |
| *** cschwede has quit IRC | 10:04 | |
| *** Masahiro has joined #openstack-swift | 10:08 | |
| *** tellesnobrega_ has joined #openstack-swift | 10:25 | |
| *** ppai has quit IRC | 10:25 | |
| *** haomaiwang has quit IRC | 10:27 | |
| *** Masahiro has quit IRC | 10:33 | |
| *** Masahiro has joined #openstack-swift | 10:34 | |
| *** tellesnobrega_ has quit IRC | 10:37 | |
| *** aix has joined #openstack-swift | 10:39 | |
| *** ppai has joined #openstack-swift | 10:39 | |
| *** tellesnobrega_ has joined #openstack-swift | 10:41 | |
| *** tellesnobrega_ has quit IRC | 10:51 | |
| *** tellesnobrega_ has joined #openstack-swift | 10:53 | |
| *** tellesnobrega_ has quit IRC | 10:54 | |
| *** Masahiro has quit IRC | 10:59 | |
| *** silor has quit IRC | 11:01 | |
| *** sandywalsh has quit IRC | 11:06 | |
| *** sandywalsh has joined #openstack-swift | 11:08 | |
| *** pberis has quit IRC | 11:19 | |
| *** pberis has joined #openstack-swift | 11:21 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Fix large out of sync out of date containers https://review.openstack.org/141019 | 11:23 |
| *** ppai has quit IRC | 11:49 | |
| *** aix has quit IRC | 11:49 | |
| *** miurahr has joined #openstack-swift | 11:54 | |
| *** kopparam has quit IRC | 11:56 | |
| *** kopparam has joined #openstack-swift | 11:57 | |
| *** Masahiro has joined #openstack-swift | 12:00 | |
| *** kopparam has quit IRC | 12:02 | |
| *** aix has joined #openstack-swift | 12:02 | |
| *** ppai has joined #openstack-swift | 12:03 | |
| *** Masahiro has quit IRC | 12:04 | |
| *** silor has joined #openstack-swift | 12:06 | |
| *** miurahr has quit IRC | 12:06 | |
| *** ho has quit IRC | 12:07 | |
| *** dmsimard_away is now known as dmsimard | 12:09 | |
| *** dmorita has quit IRC | 12:09 | |
| *** X019 has joined #openstack-swift | 12:21 | |
| *** silor1 has joined #openstack-swift | 12:22 | |
| *** silor has quit IRC | 12:23 | |
| *** oomichi has quit IRC | 12:23 | |
| *** silor has joined #openstack-swift | 12:23 | |
| *** silor1 has quit IRC | 12:26 | |
| *** sluo_afk has joined #openstack-swift | 12:31 | |
| *** sluo_laptop has quit IRC | 12:34 | |
| *** sluo_laptop has joined #openstack-swift | 12:39 | |
| *** nshaikh has quit IRC | 12:41 | |
| *** sluo_afk has quit IRC | 12:43 | |
| *** nshaikh has joined #openstack-swift | 12:43 | |
| *** kopparam has joined #openstack-swift | 12:49 | |
| *** miurahr has joined #openstack-swift | 13:12 | |
| *** bill_az has joined #openstack-swift | 13:26 | |
| openstackgerrit | Prashanth Pai proposed openstack/swift: fsync() on directories https://review.openstack.org/126923 | 13:30 |
| *** Masahiro has joined #openstack-swift | 13:48 | |
| *** rdaly2 has joined #openstack-swift | 13:50 | |
| *** Masahiro has quit IRC | 13:53 | |
| *** SkyRocknRoll has quit IRC | 13:57 | |
| *** mitz- has joined #openstack-swift | 13:58 | |
| *** miurahr has quit IRC | 14:00 | |
| *** mitz_ has quit IRC | 14:00 | |
| *** ppai has quit IRC | 14:10 | |
| *** jyoti-ranjan has quit IRC | 14:17 | |
| *** miqui_ has joined #openstack-swift | 14:21 | |
| *** miurahr has joined #openstack-swift | 14:26 | |
| *** nshaikh has left #openstack-swift | 14:35 | |
| *** kopparam has quit IRC | 14:37 | |
| *** miurahr has quit IRC | 14:43 | |
| *** rmcall has joined #openstack-swift | 14:50 | |
| *** nellysmitt has left #openstack-swift | 15:00 | |
| *** miurahr has joined #openstack-swift | 15:00 | |
| *** lpabon has joined #openstack-swift | 15:00 | |
| peluse | FYI, brief we did w/SwiftStack to coincide with Juno release on storage policies: https://www-ssl.intel.com/content/dam/www/public/us/en/documents/solution-briefs/openstack-swift-v2-solution-brief.pdf | 15:03 |
| acoles | peluse: nice | 15:05 |
| peluse | better late than never :) | 15:07 |
| portante | peluse: nice | 15:10 |
| portante | looks good, seems well written, if you take my 4th grade literacy for anything of worth | 15:10 |
| *** rmcall has quit IRC | 15:11 | |
| *** lcurtis has joined #openstack-swift | 15:13 | |
| *** rdaly2 has quit IRC | 15:17 | |
| *** tdasilva has joined #openstack-swift | 15:18 | |
| *** miurahr has quit IRC | 15:19 | |
| peluse | portante, heh, thanks. Apparently our marketing folks are only skilled in powerpoint so we had help from a third party with actually putting it together... | 15:21 |
| *** occupant has quit IRC | 15:25 | |
| *** SkyRocknRoll has joined #openstack-swift | 15:28 | |
| *** tellesnobrega has quit IRC | 15:35 | |
| *** Masahiro has joined #openstack-swift | 15:37 | |
| *** rdaly2 has joined #openstack-swift | 15:39 | |
| *** Masahiro has quit IRC | 15:41 | |
| *** rdaly2 has quit IRC | 15:44 | |
| *** silor has joined #openstack-swift | 15:46 | |
| *** MooingLemur has quit IRC | 15:49 | |
| *** exploreshaifali has joined #openstack-swift | 15:49 | |
| *** MooingLemur has joined #openstack-swift | 15:49 | |
| *** SkyRocknRoll has quit IRC | 15:50 | |
| *** mahatic has joined #openstack-swift | 15:54 | |
| *** rmcall has joined #openstack-swift | 15:57 | |
| *** rdaly2 has joined #openstack-swift | 16:01 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:02 | |
| *** nshaikh has joined #openstack-swift | 16:09 | |
| *** flwang has quit IRC | 16:20 | |
| *** SkyRocknRoll has quit IRC | 16:23 | |
| *** rmcall has quit IRC | 16:36 | |
| *** kopparam has joined #openstack-swift | 16:42 | |
| *** linuxgeek_ has quit IRC | 16:43 | |
| *** david-lyle_afk is now known as david-lyle | 16:46 | |
| *** wer has quit IRC | 16:47 | |
| *** wer has joined #openstack-swift | 16:48 | |
| *** ppai has joined #openstack-swift | 16:50 | |
| *** kopparam has quit IRC | 16:55 | |
| *** Nadeem_ has joined #openstack-swift | 17:02 | |
| *** lcurtis has quit IRC | 17:02 | |
| *** nshaikh has quit IRC | 17:06 | |
| *** rledisez has quit IRC | 17:08 | |
| *** gyee has joined #openstack-swift | 17:11 | |
| *** aix has quit IRC | 17:12 | |
| *** kopparam has joined #openstack-swift | 17:20 | |
| *** Masahiro has joined #openstack-swift | 17:26 | |
| *** silor has quit IRC | 17:26 | |
| *** Masahiro has quit IRC | 17:31 | |
| *** jordanP has quit IRC | 17:32 | |
| *** rmcall has joined #openstack-swift | 17:39 | |
| *** tellesnobrega has joined #openstack-swift | 17:46 | |
| notmyname | good morning | 17:50 |
| *** gyee has quit IRC | 17:55 | |
| *** lpabon has quit IRC | 17:56 | |
| openstackgerrit | Merged openstack/swift-specs: Workflow documentation is now in infra-manual https://review.openstack.org/139394 | 17:57 |
| notmyname | I'm abandoning the stuff on http://abandoner.oliver.net.au/abandoned_changes.html | 18:00 |
| *** jistr has quit IRC | 18:00 | |
| *** ppai has quit IRC | 18:08 | |
| *** lcurtis has joined #openstack-swift | 18:12 | |
| *** mahatic has quit IRC | 18:18 | |
| *** geaaru has quit IRC | 18:20 | |
| *** gyee has joined #openstack-swift | 18:22 | |
| *** ppai has joined #openstack-swift | 18:22 | |
| clayg | heyck yeah - abandon all the things! | 18:25 |
| clayg | abandon all hope! | 18:25 |
| clayg | stormpocalypse is upon us! | 18:26 |
| notmyname | Rain of Terror is my favorite | 18:27 |
| *** exploreshaifali has quit IRC | 18:27 | |
| notmyname | for those of you playing along at home, it's raining in northern california and san francisco basically shut down today | 18:30 |
| *** ppai has quit IRC | 18:32 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Warn if multiple replicas are stored within same region/zone/node https://review.openstack.org/140478 | 18:33 |
| *** mahatic has joined #openstack-swift | 18:34 | |
| *** kopparam has quit IRC | 18:46 | |
| *** acoles is now known as acoles_away | 18:46 | |
| *** kopparam has joined #openstack-swift | 18:46 | |
| *** cschwede has joined #openstack-swift | 18:46 | |
| *** kopparam has quit IRC | 18:47 | |
| *** ChanServ sets mode: +v cschwede | 18:47 | |
| portante | notmyname: no sympathy from the NE. :P | 18:53 |
| portante | =) | 18:53 |
| notmyname | portante: in our defense, it hasn't rained here in like 2 years | 18:53 |
| portante | well that makes it worse | 18:53 |
| notmyname | but it does seem to have been somewhat blown out of proportion (eg all the schools closed today) | 18:53 |
| portante | now all the red woods are going to grow faster, right? | 18:54 |
| notmyname | I have heard reports that there is some flooding and some power out | 18:54 |
| portante | better safe than sorry | 18:54 |
| portante | hopefully all this will be blown out of proporation any the worse that happens is everybody blames the weather people for not being good at their job | 18:57 |
| portante | s/any/and/ | 18:57 |
| portante | so how is everybody's storage redundancy in SF? :) | 18:58 |
| *** lcurtis has quit IRC | 18:58 | |
| openstackgerrit | Jiri Suchomel proposed openstack/python-swiftclient: Print info message about incorrect --totals usage when neither -l nor --lh is provided. Added test coverage for --totals. https://review.openstack.org/130463 | 18:59 |
| *** flwang has joined #openstack-swift | 18:59 | |
| ctennis | power is still online on brannan st | 19:00 |
| * peluse wonders what "rain" is | 19:01 | |
| notmyname | basically this https://twitter.com/artomatik/status/543097471514472448 | 19:02 |
| *** nellysmitt has joined #openstack-swift | 19:07 | |
| *** Masahiro has joined #openstack-swift | 19:15 | |
| *** Masahiro has quit IRC | 19:19 | |
| peluse | nice pic! | 19:20 |
| cschwede | notmyname: swifterdarrell let me know if i should change anything in https://review.openstack.org/#/c/140478/ | 19:27 |
| cschwede | notmyname: swifterdarrell: also, i’m at a customer site currently and travelling back tomorrow, so please feel free to modify the patch if you don’t get a response from my side within 12hrs | 19:28 |
| notmyname | cschwede: ya we were just talking about it with clayg too | 19:33 |
| notmyname | clayg: sounds like you want something a lot more than the warning to land in this patch. or are you ok with just the warning being printed | 19:34 |
| clayg | notmyname: no, i guess just the warning is fine since we're on the release schedule but yeah I have some diffs i'm hacking on around this | 19:38 |
| clayg | notmyname: when did you want to cut the release? | 19:38 |
| notmyname | between today and monday. | 19:39 |
| notmyname | clayg: of course if the current patch conflicts (or paints into a corner) what your and/or swifterdarrell are working on, let's not do that | 19:43 |
| clayg | i'm not sure, i haven't looked at cschwede's patch that closely... none of this is terribly surprising to me i guess is the thing, everyone is like "we need to raise a warning when the ring is screwed" and I'm all like "yeah you can tell that by looking at the weights - you can either balance that or you can't unique-as-possible it's math" | 19:47 |
| clayg | so what I'm working on is to bouble out the unique-as-possible-ness as a new kind of balance, so you can be dead to rights on the weight and be way out of whack on unique-as-possible and that just comes out like another balance except instead of "wait and rebalance and repush" it's "continue adding weight to the new tier reblance and repush" | 19:48 |
| swifterdarrell | clayg: if you auto-munge weights in a way that's divorced from actual TB of storage capacity, you're then opening yourself up to filling a subset of disks first/faster (i.e. having a problem in a different place) | 19:49 |
| cschwede | swifterdarrell: which is similar to the original behaviour in Swift < 2.1, right? for example one region with a weight of 1000, another with 10? but both get the same amount of partitions | 19:51 |
| swifterdarrell | cschwede: ya, that was a possibility you could get yourself into pre Swift 2.1... severely imbalanced storage capacity could lead to smaller devices filling up faster than if partitions were spread "better" according to storage capacity | 19:53 |
| notmyname | combined with the other problem cschwede was seeing which is adding a tier and having a huge number of parts move | 19:54 |
| clayg | yes all of that - this is working as intended, instead of unbalanced rings that blasted parts all over your tiers you get to perflectly balance yourself right into availability risk (which was the whole point of the change) | 19:56 |
| clayg | so I think - depending on what you're doing - you are going to want to have balance on one or the other - if you've finished a topology migration and you're unique-as-possible balance is still out of whack then you need to "artifically" change the weights on something in a tier somewhere - which I agree should be represented as "imbalance" in data capacity (just like pre 2.1) | 19:58 |
| *** mvenesio has joined #openstack-swift | 20:01 | |
| cschwede | i was even thinking of adding a switch, making pre 2.1 the default and only taking weights into account if a user enables this. from my point of view the prio is like this: 1. ensure everything is distributed across several failure domains 2. ensure adding new tiers doesn’t flood your new servers with data 3. ensure that devices are not over-commited | 20:05 |
| cschwede | 2. + 3. will basically halt your cluster, but data is still distributed well | 20:06 |
| clayg | idk I think 1 is almost kind of a footgun for gradual capacity changes, and also it's already doing the best it can given the capacity available - #3 isn't that dangerous unless you have a big cluster - which is exactly the case you want #1 not to be a big footgun | 20:12 |
| *** tab____ has joined #openstack-swift | 20:17 | |
| *** rdaly2 has quit IRC | 20:20 | |
| *** exploreshaifali has joined #openstack-swift | 20:22 | |
| *** nellysmitt has quit IRC | 20:24 | |
| *** fifieldt_ has quit IRC | 20:32 | |
| *** mvenesio has quit IRC | 20:35 | |
| cschwede | hmm, yes. so basically the operator needs to decide what he wants | 20:35 |
| openstackgerrit | Jiri Suchomel proposed openstack/python-swiftclient: Print info message about incorrect --totals usage when neither -l nor --lh is provided. Added test coverage for --totals. https://review.openstack.org/130463 | 20:35 |
| *** Nadeem_ has quit IRC | 20:37 | |
| *** mvenesio has joined #openstack-swift | 20:41 | |
| openstackgerrit | paul luse proposed openstack/swift: New hashes.pkl format and several other reconstructor related changes https://review.openstack.org/131872 | 20:42 |
| *** fifieldt_ has joined #openstack-swift | 20:49 | |
| swifterdarrell | cschwede: clayg: re: "the operator needs to decide what he wants"; knowing what you want is the second hardest part; being happy with what you got instead is slightly harder | 21:02 |
| swifterdarrell | cschwede: clayg (that's a little lighthearted philosophical joke, not actually a comment about this builder stuff) | 21:03 |
| *** Masahiro has joined #openstack-swift | 21:03 | |
| *** Masahiro has quit IRC | 21:08 | |
| * notmyname grumbles at super long gerrit links | 21:09 | |
| clayg | eh | 21:09 |
| mattoliverau | Morning all | 21:09 |
| clayg | cschwede: i'm not entirely sure the operator needs to decide, I think I'd like to give it a shot where there's a slide that will imbalance capacity up to X % in order to solve unique as possible, and the default be like 10% or something, and then the warning is like you can either add weight *right here at this tier that's not properly provisioned* (may have been your plan already) or you can put that off and crank up the im | 21:12 |
| clayg | where is "add weight" is a) increase actual capacity, or be continue to gradually increase the weights of the existing capacity | 21:14 |
| *** ChanServ changes topic to "Swift 2.2.0 https://launchpad.net/swift/juno/2.2.0 | Priority Reviews: https://wiki.openstack.org/wiki/Swift/PriorityReviews | Ideas: https://wiki.openstack.org/wiki/Swift/ideas | Review Dashboard: http://goo.gl/r2mxbe" | 21:16 | |
| openstackgerrit | Jiri Suchomel proposed openstack/python-swiftclient: Print info message about incorrect --totals usage when neither -l nor --lh is provided. Added test coverage for --totals. https://review.openstack.org/130463 | 21:19 |
| torgomatic | clayg: you could implement such a thing as a fudge factor on the parts_wanted computation... if you're willing to let things get out of whack by 10%, multiply each drive's parts_wanted by 1.1 prior to rebalancing | 21:19 |
| torgomatic | that way, any given disk will take up to 110% of what its normal share would be prior to being considered full, so you get better dispersion that way | 21:20 |
| torgomatic | maybe | 21:20 |
| * torgomatic is spitballing here | 21:20 | |
| clayg | torgomatic: that's the ticket! | 21:23 |
| swifterdarrell | torgomatic: clayg: +1 | 21:23 |
| cschwede | clayg: torgomatic: yeah, sounds like a nice idea to allow imbalance by 10%. +1! | 21:24 |
| *** tellesnobrega_ has joined #openstack-swift | 21:24 | |
| torgomatic | great! now someone go code it ;) | 21:24 |
| clayg | torgomatic: i'm still trying to get my head around the partition dance patch | 21:25 |
| notmyname | here's a newer review dashboard: http://goo.gl/r2mxbe. also here's a dashboard for "health of swift community" http://goo.gl/JL65l7 | 21:25 |
| *** tkay has joined #openstack-swift | 21:32 | |
| *** EmilienM is now known as EmilienM|pto | 21:39 | |
| *** tkay has left #openstack-swift | 21:43 | |
| *** tab____ has quit IRC | 21:50 | |
| *** exploreshaifali has quit IRC | 21:58 | |
| *** exploreshaifali has joined #openstack-swift | 22:01 | |
| *** nellysmitt has joined #openstack-swift | 22:24 | |
| *** mvenesio has quit IRC | 22:25 | |
| *** tab has joined #openstack-swift | 22:26 | |
| *** tab is now known as Guest97 | 22:26 | |
| *** nellysmitt has quit IRC | 22:30 | |
| notmyname | mattoliverau: dfg: redbo: acoles_away: I'd love your input on a few of the trickier patches that have come up this week | 22:33 |
| notmyname | https://review.openstack.org/#/c/141019/ <- fix large out of sync containers | 22:33 |
| notmyname | https://review.openstack.org/#/c/140478/ <- warn if multiple replicas are stored in the same tier | 22:33 |
| clayg | glange_: ^ he ment to include you too | 22:34 |
| notmyname | https://review.openstack.org/#/c/140879/ <- only move too-close-together replicas when they can spread out | 22:34 |
| clayg | ^ ohhhh that one is tricky | 22:34 |
| notmyname | portante: we could probably use a pinch hitter on these too ;-) | 22:34 |
| *** kopparam has joined #openstack-swift | 22:37 | |
| clayg | torgomatic: I wonder if I should try running that db replicator test against the container backend code from 1fcd6096^ | 22:38 |
| mattoliverau | notmyname: I'll review them all today :) | 22:40 |
| notmyname | ...he say's before diving in to them ;-) | 22:40 |
| notmyname | says* | 22:40 |
| mattoliverau | ok, I'll start reviewing them today, and see how far i get :P | 22:41 |
| notmyname | heh | 22:41 |
| notmyname | thanks! | 22:41 |
| notmyname | the 2nd one (the warning one) is shortest. but has probably had the most debate (hello bikeshed.org?) | 22:41 |
| mattoliverau | lol, so I've noticed in scroll back :P | 22:42 |
| torgomatic | clayg: sounds like a good idea to me | 22:43 |
| *** gyee has quit IRC | 22:45 | |
| *** dmsimard is now known as dmsimard_away | 22:49 | |
| *** Masahiro has joined #openstack-swift | 22:52 | |
| *** Guest97 has quit IRC | 22:55 | |
| *** Masahiro has quit IRC | 22:57 | |
| *** kopparam has quit IRC | 23:09 | |
| clayg | c | 23:09 |
| clayg | hrm... this is not the pdb i was looking for | 23:10 |
| notmyname | torgomatic: parts_wanted is 0 if the weight for that drive is 0, right? eg if it's an initial ingest or a final step of a drain or just left in the cluster | 23:11 |
| *** sungju has joined #openstack-swift | 23:28 | |
| *** tellesnobrega has quit IRC | 23:30 | |
| *** mahatic has quit IRC | 23:37 | |
| *** mahatic has joined #openstack-swift | 23:37 | |
| *** tellesnobrega has joined #openstack-swift | 23:42 | |
| torgomatic | notmyname: yes | 23:48 |
| notmyname | torgomatic: thanks | 23:48 |
| *** ahonda has quit IRC | 23:55 | |
| *** tdasilva has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!