From 8acbec0613ea1dd54c500ca81fd83725f4025440 Mon Sep 17 00:00:00 2001 From: Hans van den Bogert Date: Mon, 28 Aug 2017 23:05:38 +0200 Subject: [PATCH] [doc] Remedy falsehood about losing up to 2/3 of a cluster First, this depends on the MIN_SIZE parameter of a pool. Second even if the MIN_SIZE is one this would only hold up to six node cluster. Imagine a nine node cluster; losing two-thirds of a cluster can lead to a situation where all replicas of an object are offline. I.e., you can only lose up to two failure domains, in this case nodes, when you have a replication size of three. --- Documentation/advanced-configuration.md | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/Documentation/advanced-configuration.md b/Documentation/advanced-configuration.md index c95e6ca59..122f1cf39 100644 --- a/Documentation/advanced-configuration.md +++ b/Documentation/advanced-configuration.md @@ -311,9 +311,10 @@ Now if you run `ceph -s` or `rookctl status` you may see "recovery" operations a PGs in "undersized" and other "unclean" states. The cluster is essentially fixing itself since the number of replicas has been increased, and should go back to "active/clean" state shortly, after data has been replicated between -hosts. When that's done you will be able to lose 2/3 of your storage nodes and -still have access to all your data in that pool. Of course you will only have -1/3 the capacity as a tradeoff. +hosts. When that's done you will be able to lose two of your storage nodes and +still have access to all your data in that pool, since the CRUSH algorithm will +guarantee that at least one replica will still be available on another storage node. +Of course you will only have 1/3 the capacity as a tradeoff. ### Setting PG Count