From: Gregory Farnum Date: Wed, 26 Jun 2019 15:14:11 +0000 (-0700) Subject: doc: update erasure code recovery description X-Git-Tag: v15.1.0~2330^2~1 X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=4f246cf74d17549f71118e8afbf60793b20ce2db;p=ceph.git doc: update erasure code recovery description Signed-off-by: Greg Farnum --- diff --git a/doc/rados/operations/erasure-code.rst b/doc/rados/operations/erasure-code.rst index a94ee5a252f..c49d06bb52e 100644 --- a/doc/rados/operations/erasure-code.rst +++ b/doc/rados/operations/erasure-code.rst @@ -165,11 +165,17 @@ More information can be found in the `cache tiering Erasure coded pool recovery --------------------------- - -Erasure coded pools could not recover when there were less than min_size -shards available. This restriction has been removed in Octopus, which means -that as long as there are enough shards available for data to be recovered, -we will allow recovery to proceed below min_size. +If an erasure coded pool loses some shards, it must recover them from the others. +This generally involves reading from the remaining shards, reconstructing the data, and +writing it to the new peer. +In Octopus, erasure coded pools can recover as long as there are at least *K* shards +available. (With fewer than *K* shards, you have actually lost data!) + +Prior to Octopus, erasure coded pools required at least *min_size* shards to be +available. (We generally recommend min_size be *K+2* or more to prevent loss of writes and data.) +This conservative decision was made out of an abundance of caution when designing the new pool +mode but also meant pools with lost OSDs but no data loss were unable to recover and go active +without manual intervention to change the *min_size*. Glossary --------