After catch ASK send_cluster_command shouldn't update slots->node cache
Hailei opened this issue · 1 comments
Hailei commented
After catch ASK send_cluster_command shouldn't update slots->node cache. it will result in more redirection.
if errv[0] == "MOVED" || errv[0] == "ASK"
if errv[0] == "ASK"
asking = true
else
# Serve replied with MOVED. It's better for us to
# ask for CLUSTER NODES the next time.
@refresh_table_asap = true
end
newslot = errv[1].to_i
node_ip,node_port = errv[2].split(":")
@slots[newslot] = {:host => node_ip,
:port => node_port.to_i}
else
raise e
end
antirez commented
Thanks! Fixed.