David Ribeiro Alves has posted comments on this change. Change subject: WIP: Kudu Consistency Blog Post Pt1 ......................................................................
Patch Set 3: (35 comments) http://gerrit.cloudera.org:8080/#/c/7019/3/_posts/2017-05-30-kudu-consistency-pt1.md File _posts/2017-05-30-kudu-consistency-pt1.md: PS3, Line 3: Pt1 > nit: Part 1 Done PS3, Line 6: multi-part short blog post series > nit: series of short blog posts (multi-part implied by "series") Done PS3, Line 7: it’s > nit: its Done PS3, Line 7: and > nit: , and Done PS3, Line 10: follow up > follow-up Done Line 11: of how they all play together and what are our future plans. > would be nice to also include what is discussed in Part 1. Done PS3, Line 15: : > nit: ? instead of : Done PS3, Line 17: users and requests : are executed by many different machines, > Howabout: "...users while coordinating requests across many machines" had to choose one, chose Will's PS3, Line 20: how Kudu implements this > "how these systems implement" Done PS3, Line 21: , > remove Done PS3, Line 24: making writing correct applications against it easy to code and reason about > Howabout: "making correct applications easy to code and reason about" or "m Done PS3, Line 27: > nit: an extra space Done PS3, Line 29: single thread > "single-threaded" Done PS3, Line 29: , > nit: remove Done PS3, Line 29: single machine > single-machine Done PS3, Line 31: are designed to be distributed, : which means that there are many more things that can fail, and must take store and process data : on many machines/cores, which means that many things happen at the same time in different locations > howabout: "are designed to be distributed, storing and processing data acro Done PS3, Line 34: with > remove Done PS3, Line 34: > with Done PS3, Line 35: > nit: I prefer a comma here to split up the two compound adjectives Done PS3, Line 35: single thread > single-threaded Done PS3, Line 35: > - Done PS3, Line 36: where failures are common and many things happen at the same time > nit: surround with ,'s Done PS3, Line 38: > nit: an extra space Done PS3, Line 42: ref aphyr’s post > consider adding the actual link Done PS3, Line 43: , > ; Done Line 45: ## Motivation behind the consistency design decisions in Kudu: > I like the first paragraph as an introduction to design tradeoffs, even kno Chose to re-title. Andrew let me know if it makes more sense now. PS3, Line 49: a term from machine learning > I thought "no such thing as a free lunch" came from economics Done PS3, Line 49: trade offs > trade-offs or tradeoffs (consistently one or the other) Done PS3, Line 61: which is > remove Done PS3, Line 67: is > are Done PS3, Line 74: for analytics > remove Done PS3, Line 74: architecture > architectural Done PS3, Line 95: semantics > FWIW I think semantics is a good word here ended up leaving, for "consistency" :) PS3, Line 95: semantics > "guarantees", "constraints"? Done PS3, Line 101: Kudu currently lacks support for atomic multi-row mutation operations (i.e. mutation : operations to more than one row in the same or different tablets, planned as a future feature) so, : for the write path, we’ll be talking about the consistency semantics of single row mutations. : In this context we’ll discuss Kudu’s properties more from a key/value store standpoint. On the : other hand Kudu is an analytical storage engine so, for the read path, we’ll also discuss the : semantics of large (multi-row) scans. This moves the discussion more into the field of traditional : DBMSs. These ingredients make for a non-traditional discussion that is not exactly apples-to-apples : with what the reader might be familiar with, but our hope is that it still provides valuable, or : at least interesting, insight. > maybe consider leaving the write and read discussion to their respective se moved it as a "note" in the previous section -- To view, visit http://gerrit.cloudera.org:8080/7019 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Icaec0c8ace20651a65901a8e1786e785265540d1 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: gh-pages Gerrit-Owner: David Ribeiro Alves <davidral...@gmail.com> Gerrit-Reviewer: Alexey Serbin <aser...@cloudera.com> Gerrit-Reviewer: Andrew Wong <aw...@cloudera.com> Gerrit-Reviewer: David Ribeiro Alves <davidral...@gmail.com> Gerrit-Reviewer: Will Berkeley <wdberke...@gmail.com> Gerrit-HasComments: Yes