#general
- j
jolly-chef-92794
03/24/2016, 10:49 PMAnyway, I’ve gotta head out now - w
witty-crayon-22786
03/24/2016, 10:59 PManother call for assistance on https://rbcommons.com/s/twitter/r/3592/ ... anyone able to take a look at the core changes? - m
mammoth-pharmacist-54215
03/24/2016, 11:00 PMwill take a look in a moment after I finish going over your internal docs RB 😛 - m
mammoth-pharmacist-54215
03/24/2016, 11:07 PMwhats the implication for removing num_chunking_units? - w
witty-crayon-22786
03/24/2016, 11:13 PM@mammoth-pharmacist-54215: nothing, in theory... the consumer was deleted a little while ago - w
witty-crayon-22786
03/24/2016, 11:13 PM - w
witty-crayon-22786
03/24/2016, 11:14 PMwill mention that in the description - m
mammoth-pharmacist-54215
03/24/2016, 11:14 PMgreat thanks - m
mammoth-pharmacist-54215
03/24/2016, 11:15 PMI find it amusing that there is a random 55 in with all the 42’s for num_chunking_units - m
mammoth-pharmacist-54215
03/24/2016, 11:16 PMnot sure how to make RB not do a ship-it instead of fix-it and ship it but still leave a comment - m
mammoth-pharmacist-54215
03/24/2016, 11:17 PMI wouldn’t consider my comment blocking. - r
rough-minister-58256
03/24/2016, 11:28 PMdo you not see the ‘Open an Issue’ tickbox when adding comments? - r
rough-minister-58256
03/24/2016, 11:29 PMPasted image at 2016_03_24 04_29 PM.png - r
rough-minister-58256
03/24/2016, 11:29 PMyou should have a little checkbox like that that allows you to mark an issue as blocking/nonblocking - r
rough-minister-58256
03/24/2016, 11:29 PMwhich is the mechanism that controls ‘ship it’ vs ‘fix it and ship it’ when a shipit is given w/ comments - m
mammoth-pharmacist-54215
03/25/2016, 12:00 AMperfect, seems obvious enough simple smile - m
mammoth-pharmacist-54215
03/25/2016, 1:35 AM@bored-art-40741: sorry about the delay but I have up an RB about the scala-platform changes. https://rbcommons.com/s/twitter/r/3609/ Let me know if that works for you when you get a chance. - b
bored-art-40741
03/25/2016, 2:07 PMReviewed. Looks reasonable, though just by eyeballing it it's not clear that it gives a good error message - l
limited-country-20626
03/25/2016, 2:20 PMI have opened access to a couple documents related to the recent rebuild of Foursquare's artifact cache. The design of the our topology is here: https://docs.google.com/document/d/1OJ0zSGWNgPAZbIuG66UtkYoY8a0T3yfW2aVMGveyrbE/edit#heading=h.zarr45ev0mlh The nginx.conf we are using is here: https://gist.github.com/mateor/4938d1cb521b19ef889e. - l
limited-country-20626
03/25/2016, 2:20 PMOpensourcedversionofNewPantsArtifactCache.txt - l
limited-country-20626
03/25/2016, 2:21 PMThe only change from what we have up internally is the list of server names. This was designed and implemented by @happy-morning-1599, who lurks in this slack sometimes ; ) - l
limited-country-20626
03/25/2016, 2:22 PMI am going to x-post this to the mailing list as well - a
ancient-stone-50795
03/25/2016, 2:23 PM@limited-country-20626: looking at your diagram, looks like you have 3 copies of the cache. How do cached files get written to all 3 nginx serves? - a
adorable-glass-89507
03/25/2016, 2:28 PMawesome, thanks @limited-country-20626 - a
adorable-glass-89507
03/25/2016, 2:28 PM@limited-country-20626: I submitted a request for access - b
bored-art-40741
03/25/2016, 2:30 PM@ancient-stone-50795: We're going to clarify a little in the doc, but the nginx config uses consistent hashing for both PUTs and GET/HEADs - b
bored-art-40741
03/25/2016, 2:30 PMSo the whole thing is sharded - b
bored-art-40741
03/25/2016, 2:31 PMEach node is actually two layers: a thin proxy layer that redirects or proxies to the correct (hash based) node depending on the method, and then the back layer which assumes it's the authoritative server and just handles the request - l
limited-country-20626
03/25/2016, 2:32 PM@adorable-glass-89507: I blanket approved pants-devel mailing list - but I also just approved you indivdiually. - a
adorable-glass-89507
03/25/2016, 2:32 PMok thanks