[Gluster-devel] all server down

[Emmanuel Dreyfus]

Hi

on release-3.3,when all replicated subvolumes are down, FUSE receive 
EAGAIN. Is that the intended behavior? IF it is, then what is the proper
behavior for the FUSE implementation when receiving that? Should the
operation be replayed after a while? (how long?) Or should EAGAIN be 
reported to userland? What does Linux?

-- 
Emmanuel Dreyfus
manu at netbsd.org