Thanks for your reply Ralph.
In order to preserve the performance could be better to guarantee that when
the buffer gets full and the I/O takes place, the entire content of the buffer
is written on the alternative appender, instead of only the last message. This
will guarantee that all
Unfortunately, the appender has no idea how much data is being buffered as the
buffer is managed by the Socket.
Ralph
On Mar 27, 2013, at 4:11 AM, Daniele Menotti wrote:
Thanks for your reply Ralph.
In order to preserve the performance could be better to guarantee that when
the
Could we have a decorated socket that allows listeners to know its
buffer status and exposes a method to flush? Then the appender could use
that
On 3/27/2013 9:27 PM, Ralph Goers wrote:
Unfortunately, the appender has no idea how much data is being buffered as the
buffer is managed by the