From 8257056e5254e50d6b8c57c8ee1c85092f7e00a5 Mon Sep 17 00:00:00 2001 From: Chris Lu Date: Mon, 28 Dec 2020 12:25:22 -0800 Subject: [PATCH] Updated Path Specific Filer Store (markdown) --- Path-Specific-Filer-Store.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Path-Specific-Filer-Store.md b/Path-Specific-Filer-Store.md index e59d0f3..419d26f 100644 --- a/Path-Specific-Filer-Store.md +++ b/Path-Specific-Filer-Store.md @@ -6,9 +6,9 @@ In most cases, you would just need to setup one filer store. However, there are some examples that one filer store is not enough: -* A portion of the data is critical and needs strong consistency and is ok with high latency. +* A portion of the data is critical. It needs strong consistency and is ok with high latency. * With too many updates in a directory, the Cassandra with LSM tree becomes slower with too many tombstones. You may want to use Redis for that directory. -* One filer store can not handle the amount of files. We can add more file stores for big directories to ensure linear scalability. +* One filer store can not handle the total amount of files. We can add more file stores for big directories to ensure linear scalability. * One directory operations are hot. Having a separate store can physically isolate from and avoid impact to other directories. # How to add path-specific filer stores?