Hi, I'm new to the list. Apologies if this has been asked and answered before, my search didn't unearth it. We're running GlusterFS 3.2.1 on a 2-server load-balanced web-hosting cluster, to replicate web-uploaded files between the two web servers. We're using nginx. We're about to move a Drupal site to this cluster that uses the Boost static file cache. This file cache needs to be replicated. I've heard that reads from glusterfs are around 20 times slower than from ext3: http://groups.drupal.org/node/61908#comment-193143 . This post suggests a workaround of doing reads directly from the local brick. Is this a safe thing to do: do reads from the brick and writes to the glusterfs volume? Thanks, Dave
